Skip to main content

Full text of "Cosmic Rays"

See other formats



DIyUIML/ nUjjl Massachusetts Institute of Technology 

i All 


Cosmic Rays 

In these lively pages the foremost living expert 
in the field unfolds a simple yet scientifically 
accurate historical narrative of cosmic rays: 
how physicists discovered them and, after a 
half century of hard work, analyzed their com- 
position. Also how, in the process, they discov- 
ered a host of new particles, born out of energy 
and living but a minute fraction of a second, 
opening up the field of elementary-particle 
physics. Also, how the study of cosmic rays 
revealed new vistas in astrophysics and cosmo- 
logy; and how, for all their labors, physicists 
are now faced with a more challenging and 
formidable problem than any yet encountered 
— the problem of the origin of cosmic rays. 

The book is to a large extent autobiographical. 
In it the author has tried to recapture some of 
the excitement that a scientist feels when he 
ventures into an unexplored field. 

Dr. Rossi has written this book in a form that 
should be easily understandable to any edu- 
cated layman, without, however, compromising 
with scientific accuracy. The material is en- 
tirely up to date, and includes the latest results 
of space research pertinent to Cosmic Ray 

About the Author 

Bruno Rossi obtained the degree of Doctor of 
Physics from the University of Bologna, Italy, 
in 1927. From 1928 to 1932 he was Assistant 
of Physics at the University of Florence. In 
1932 he was named Professor of Physics at 
the University of Padova. Leaving Italy in 1938, 
he stayed briefly in Copenhagen as a guest at 
Professor Bohr's Institute of Theoretical Phys- 

(Continued on back flap) 




/B-fSS|K/^B Vtk Bfc^ralaftl. 

^-' MCMXCVII ^y 

Cosmic Rays 

McGraw-Hill Paperbacks in Physics 

Arthur F. Kip University of California, Berkeley 
Consulting Editor 

James J. Brophy Semiconductor Devices 

Bruno B. Rossi Cosmic Rays 

Hugh D. Young Statistical Treatment of Experimental Data 

Professor Bruno Rossi 

Department of Physics 
Massachusetts Institute of Technology 

McGraw-Hill Book Company 

New York 
San Francisco 

Cosmic Rays 

Copyright © 1964 by McGraw-Hill. Inc. All Rights Reserved. Printed in the 
United States of America. This book, or parts thereof, may not be reproduced 
in any form without permission of the publishers. 

Library of Congress Catalog Card Number 64-17570 




A steady rain of charged particles, moving at nearly the speed of 
light, falls upon our planet at all times and from all directions. 
These particles, known as cosmic rays, are just the nuclei of ordi- 
nary atoms stripped of their electrons— for the most part nuclei of 

The property of cosmic rays that sets the rays apart from all 
other kinds of radiation and accounts for the extraordinary role 
they have played in the development of modern physics is the very 
large individual energies of cosmic-ray particles. Before the dis- 
covery of cosmic rays, the particles of highest energy known to 
physicists were those emitted in the spontaneous decay of radio- 
active atoms. Some time later, largely in an effort to duplicate the 
effects of cosmic rays under controlled laboratory conditions, 
physicists began developing accelerators capable of yielding par- 
ticles of higher and higher energies. Recently, man-made machines, 
having achieved energies about 10,000 times those characteristic 
of natural radioactivity, have overtaken the "average" cosmic-ray 
particle. But cosmic rays have a very wide range of energy, and a 
few of them are known to have more than a billion times the energy 
of the particles produced by the most powerful of our accelerators. 

In this book I have tried to relate how, starting from seemingly 
trivial observations, physicists discovered the existence of cosmic 
rays; how, through half a century of hard work, they finally found 
out precisely what cosmic rays are ; how, in the process, they discov- 
ered a host of new particles that are born out of energy and live 




but a minute fraction of a second, and thus opened the field of 
elementary-particle physics; and how, for all their labors, they now 
find themselves faced with a more challenging and formidable 
problem than any they already solved: the problem of the origin of 
cosmic rays. 

I felt that this story might have some interest not only for the 
professional physicist but for any person with a curiosity about 
modern science and about the manner in which scientific discov- 
eries are made. Therefore, I have tried to keep the language as 
simple as possible without violating scientific accuracy, and I have 
omitted all the technical details and mathematical developments 
that are not needed for understanding the essential parts of the 
story. I know that in doing so I have failed to give proper recog- 
nition to many distinguished scientists who made vital contri- 
butions to the development of cosmic-ray physics. It would be 
futile for me to apologize to them; instead, I trust in their under- 
standing of the constraints imposed by the character of this book. 

I must also warn the reader that because of my personal involve- 
ment in the subject, I could not be expected to write the story of 
cosmic rays as impartially as I might have had I been merely a 
spectator. In fact, what I have told in this book is to a large extent 
the story of how my understanding of cosmic-ray effects developed 
through the years. Therefore, it is quite possible that my own work 
and that of my collaborators may have been given undue promi- 

Many friends have generously helped with advice and sugges- 
tions. I am deeply grateful to all of them, but particularly to 
Giuseppe Occhialini and the late Francis Friedman, who took the 
time to read the whole manuscript and did not spare me their 
constructive criticism. I wish also to express my appreciation to 
Anthony Wiggenhorn for his most conscientious and competent 
editorial work. 

Bruno Rossi 


Preface vii 

1 A radiation from outer space 1 

2 Atoms and radiations 13 

3 The nature of cosmic rays 27 

4 Puzzling clues 43 

5 Using the earth as a magnet to analyze 
cosmic rays 53 

6 Positrons and the materialization 
of energy 77 

7 Electrons, photons, and showers 87 

8 Mil mesons 101 

9 Pi mesons 125 

10 Nuclear interactions of cosmic rays 145 

1 1 What cosmic rays are and what 
they do in the atmosphere 163 

12 Giant showers of the atmosphere 179 

13 The Van Allen radiation belt 193 

1 4 Cosmic rays and the sun 205 

15 The origin of cosmic rays: 
an unsolved problem 219 


A Mass per unit area 233 

B Powers of ten 235 

C Logarithmic scales 237 

D Energy and momentum 239 

E The electron volt 243 

F Computation of magnetic rigidity 245 

G The neutron and the structure of 
atomic nuclei 247 

H The neutrino 251 

I Elementary particles 255 

Index 261 

The subject [of cosmic rays] is unique in 
modern physics for the minuteness of the 
phenomena, the delicacy of the observations, 
the adventurous excursions of the observers, 
the subtlety of the analysis, and the grandeur 
of the inferences. 

Karl K. Darrow 

A radiation 
from outer space 


At six o'clock on the morning of August 7, 1912, a balloon ascended 
from a field near the town of Aussig, in Austria. In the gondola of 
the balloon were three men: a navigator, a meteorologist, and a 
physicist. During the next 2]4 hours, the balloon rose to an 
altitude of 13,000 feet while drifting rapidly northward. For 
another hour it floated between 13,000 and 16,000 feet. At noon the 
balloon touched down near the German town of Pieskow, 30 miles 
east of Berlin and some 125 miles from Aussig. 

The physicist and leader of the flight was Victor F. Hess. He 
had taken with him three electroscopes of the kind then being used 
to detect and measure the radiation emitted by radium and other 
radioactive substances. While his companions took care of the 
navigation and measured altitude and temperature, Hess watched 
his instruments and recorded their readings. A few months later, 
after a careful study of the data, he presented to the scientific 
community a conclusion of far-reaching significance. In the 



Cosmic Rays 

November, 1912, issue of the German journal Physikalische Zeil- 
schrifl, Hess summarized his work with the statement: "The results 
of my observations are best explained by the assumption that a 
radiation of very great penetrating power enters our atmosphere 
from above." 

This was the beginning of one of the most extraordinary 
adventures in the history of science. Subsequent investigations of 
the "radiation from above" opened up the new and bewildering 
world of high-energy physics. Here scientists found particles of 
subatomic dimensions, with energies thousands, millions, billions, 
trillions of times greater than the energy of particles emitted by 
radioactive materials found on earth. Here for the first time they 
witnessed processes in which particles of matter are created out of 
energy and then promptly disappear in giving birth to other 
particles. Beyond this, Hess's discovery revealed new vistas in 
astrophysics and cosmology. The mysterious radiation was found 
to carry important messages concerning the physical conditions of 
the distant regions of space through which it had traveled on its 
way to the earth. And finally, in an effort to explain the origin of 
the radiation, physicists developed a number of novel ideas about 
the nature of the events that take place in stars and in the masses 
of dilute gas that fill interstellar space. 

But in August, 1912, no one — including Hess — had any 
reason to suspect such developments. What then, had prompted 
Hess to perform his experiment? And on what evidence did he base 
his conclusion? 

Most scientific discoveries sprout from seeds that have been 
lying about for a long time. Hess's discovery was no exception. 
For more than a century physicists had known that a gold-leaf 
electroscope (Fig. 1-1) or similar instrument will not hold an 
electric charge indefinitely. The phenomenon of loss of charge is 
familiar to anyone who has ever taken a high school physics course. 
When a charged piece of glass, for example, is touched against the 
metal rod of an electroscope, the two leaves suspended from the 
rod fly apart. This happens because both leaves have acquired the 

A radiation from outer space 3 

same charge (positive in this case) and like charges repel each other. 
The leaves remain apart for a time after the glass has been re- 
moved; but eventually, little by little, they drop to their original 
positions as the electric charge gradually disappears. 

In an ordinary electroscope the charge escapes mainly through 
the insulating sleeve that separates the rod from the metal case. 
With special care, this leak can be reduced practically to zero; but 
even so, the electroscope will not retain its charge indefinitely. Nor 
will the use of any particular gas (instead of air) in an airtight 
electroscope case affect the final result. 

By the end of the nineteenth century, physicists had learned 
enough about the structure of matter to be able to explain the 
spontaneous discharge of electroscopes. Briefly, they knew that 
matter consists of atoms, that each chemical element represents a 
distinct type of atom, and that the atoms of one or more elements 
join in various combinations to form molecules of different chemical 
substances. While their picture of what atoms "look like" was far 
from precise, they knew that atoms ordinarily contain equal 
quantities of positive and negative charge. To this overall balance 

Fig. 1-1 Gold-leaf electroscope. The 
two gold leaves hang from a rnetal rod 
that is separated from the metal rase of 
the electroscope by an insulating sleeve. 
The leaves are viewed through a glass 
window. Usually the case is electrically 

4 Cosmic Rays 

of opposite charges they ascribed the neutral state of atoms and 
molecules. Moreover, they knew that the charges themselves 
possess a "granular structure," in other words, that they occur as 
simple multiples of some indivisible elementary charge. Indeed, 
in 1897 the English physicist J. J. Thomson had discovered a 
particle of very small mass — the electron — which he identified as 
the fundamental unit of negative charge. 

Now according to this picture, the discharge of the electro- 
scope was to be explained by the fact that gas molecules occasion- 
ally lose their electrical neutrality. \ molecule may lose an electron 
and thus be left with an excess positive charge. The electron either 
remains free or attaches itself to a neutral molecule. A number of 
experiments had shown conclusively that the air or gas surrounding 
the electroscope leaves is always slightly ionized, that is, always 
contains a small percentage of "free" electrons and charged mole- 
cules, or ions as they are known. Thus, if the leaves are positively 
charged, they attract negative ions and repel positive ions (Fig. 
1-2). The positive charge that holds the leaves apart is gradually 
neutralized by negative ions, and the leaves drop. Conversely, 
negatively charged leaves attract positive ions with the same result. 

A radiation from outer space 


Fig. 1-2 Discharge of an electroscope by 
gas ions. 

But what was the reason for the ever-present, if slight, ioniza- 
tion of gases? Physicists seeking to answer that question in the 
early years of the twentieth century were aware that various kinds 
of radiation could ionize gases. In 1895 Wilhelm Roentgen had 
discovered X-rays and had found that an electroscope discharged 
immediately when placed in front of an X-ray tube. The next year 
Henri Becquerel discovered radioactivity, and in 1898 Pierre and 
Marie Curie succeeded in isolating radium. This element and other 
radioactive substances also emitted ionizing rays capable of dis- 
charging an electroscope. In the light of these facts, most physicists 
were inclined to consider the observed ionization of gases as result- 
ing from some sort of weak radiation rather than from a spon- 
taneous breakup of gas molecules. 

One obvious possibility was that the radiation came from small 
traces of radioactive substances present in the materials from which 
electroscopes were made. This guess turned out to be partly correct; 
for some of the ionization was often traced to such impurities. 
However, radioactive contamination of the instrument could not 
account for the whole effect. Experimenters were able to decrease 
the rate of discharge by surrounding an electroscope with lead or 
water; thus part of the ionizing radiation had to come from outside. 
At the same time, the thick layers of matter required to produce 
an appreciable decrease pointed to a radiation of considerable 
penetrating power. 

For a decade the prevailing opinion attributed the source of 
the penetrating radiation to minute quantities of radioactive 
materials in the earth's crust. Clearly, one could test the assump- 
tion by comparing observations at different altitudes. If the 
radiation responsible for the discharge of the electroscope came 
from the earth, it should be strongest near the earth's surface and 
become progressively weaker with increasing altitude. 

Several investigators, including Thomas Wulf in Germany and 
A. Gockel in Switzerland, undertook to put the hypothesis to an 
experimental test. By that time, electroscopes were much more 
refined and reliable than the old-fashioned gold-leaf type. Par- 

6 Cosmic Rays 

ticularly suitable as a radiation meter was the instrument designed 
by Wulf in 1909. He replaced the gold leaves with two very thin 
metal wires held under tension by a light quartz Gber (Fig. 1-3). 
When charged, the two wires would repel each other, and the 
separation could be measured by means of a microscope. In 1910 
Wulf carried one of his electroscopes to the top of the Eiffel Tower; 
in 1912 Gockel used a similar instrument in a balloon ascent. 
Neither found what they had expected. The rate of discharge did 
not decrease with altitude, or at least did not decrease as fast as 
they had anticipated. 

This was the situation when Hess became interested in the 
problem and initiated a series of balloon flights that culminated in 
the memorable ascent of August, 1912. His electroscopes were of 
the Wulf type, constructed with particular care. Two of the three 
instruments had airtight cases so that the internal air pressure 
remained constant at all altitudes. Consequently, the sensitivity of 
these instruments was independent of altitude. (The sensitivity 

Fig. 1-3 The electroscope that was 
developed by Wulf in 1909 and was 
used in many of the early experi- 
ments on cosmic rays. 

.4 radiation from outer space 

of a uonpressurized electroscope decreases with increasing altitude, 
because the number of ion pairs produced in a given volume of gas 
by a given intensity of radiation is proportional to the gas pressure. 
Ions are always produced — and neutralized — by twos: one posi- 
tive and one negative. Hence the term ion pair.) 

As the balloon began its ascent through the atmosphere, Hess 
found that the ionization became somewhat weaker at first, as 
indicated by a slower rate of electroscope discharge. Unquestion- 
ably, there was a radiation emanating from the earth's crust. But 
above 2,000 feet the trend reversed itself and the ionization began 
to increase gradually with altitude, as though the balloon were 
moving toward the source of the ionizing radiation instead of away 
from it. Indeed, at 16,000 feet the electroscopes were discharging 
about four times faster than they had at ground level. It was in 
order to explain this unexpected increase that Hess postulated a 
radiation falling upon the earth from somewhere beyond the 

His assumption was certainly a bold one, and many years were 
to pass before it became generally accepted. First of all, other 
experimenters had to check the findings reported by Hess. When 
they did, they found that the increase of radiation strength with 
altitude continued well above 16,000 feet. Especially noteworthy 
were the daring balloon flights carried out by W. Kohlhbrster of 
Germany between 1913 and 1919. In his flights he reached a maxi- 
mum altitude of about 28,000 feet, where he found an ionizing 
radiation considerably stronger than that detected by Hess. 

Second, it was necessary to explore all other possible — and 
less revolutionary — explanations for the observed ionization. 
C. T. R. Wilson, the inventor of the cloud chamber and one of the 
leading experts in ionization phenomena, suggested that the radia- 
tion was produced by thunderstorms high up in the atmosphere. 
Other physicists thought the atmosphere might contain small traces 
of radioactive gases. (Some radioactive substances, such as radon, 
element 86, were known to exist in the gaseous state.) If, for some 
reason, these radioactive gases had a tendency to concentrate in the 


Cosmic Rays 

upper layers of the atmosphere, they would explain the increased 
ionization at higher altitudes. 

On both of these assumptions the intensity of the unknown 
radiation should vary according to weather conditions, it should 
change with the hour, day, and season. For thunderstorms ob- 
viously did not occur at all times, and it was hardly conceivable 
that the distribution of the hypothetical radioactive gases in the 
atmosphere would remain constant at all times of the day or year 
and through all changes of weather. Now, Hess had already pointed 
out the apparent absence of such intensity variations, and later 
investigations confirmed his findings. Clearly, despite disagree- 
ments and uncertainties due to the limited accuracy of the instru- 
ments in use at the time, the radiation was on the whole remarkably 
uniform. It came by day and by night, in summer and winter, in 
rain or shine, with little change from one day to the next, and with 
little change from one place to another at the same altitude. 

The work of Millikan and Regener 

Still, not all physicists were satisfied. Most skeptical among the 
skeptics was Robert A. Millikan, professor of physics at the Cali- 
fornia Institute of Technology. He and his collaborators decided to 
determine for themselves whether the experimental data reported 
by Hess and other workers were correct, whether there actually 
were compelling reasons to believe in the existence of a radiation 
from outer space. The result was a series of remarkable experiments 
carried out between 1923 and 1926, involving measurements made 
under water as well as at high altitudes. Millikan 's experiments 
convinced himself and nearly everyone else in the scientific com- 
munity that the radiation discovered by Hess did come from 
beyond the earth's atmosphere. And it was Millikan who gave the 
name cosmic rays to this radiation. 

Millikan 's last doubts about the existence of cosmic rays were 
apparently dispelled by an experiment performed at Muir Lake 

A radiation from outer space 

and Arrowhead Lake in the San Bernardino Range of southern 
California. The respective altitudes of the lakes are 11,800 and 
5,100 feet. To any radiation passing vertically through the atmos- 
phere, the additional 6,700 feet of air above Arrowhead Lake 
would represent an absorbing mass per unit area of lake surface 
equivalent to 6 feet of water. By sinking his electroscope to various 
depths in the two lakes, Millikan found that "within the limits of 
observational error, every reading in Arrowhead Lake corre- 
sponded to a reading six feet farther down in Muir Lake, thus 
showing that the rays do come in definitely from above and that 
their origin is entirely outside the layer of atmosphere between the 
levels of the two lakes." 

A more detailed explanation of the argument on which Milli- 
kan based his conclusion appears in Fig. 1-4. As a footnote, it 
should be pointed out that his argument is neither foolproof nor 
completely correct. For reasons discussed in Chap. 8, equal masses 
per unit area 1 of air and water do not absorb exactly the same 
fraction of cosmic rays. Had the experimental data been more 
accurate than the state of the art permitted at the time, Millikan 
would have detected a difference between readings in Arrowhead 
Lake and readings 6 feet farther down in Muir Lake. Conceivably 
this finding might have prevented him from reaching the correct 

The work of Millikan's team made history not only because of 
the scientific results obtained but also because of the novel and 
ingenious techniques employed. An important innovation was the 
application of sounding balloons to cosmic-ray research. Hess and 
Kohlhorster had had to accompany their electroscopes in order to 
observe them. The use of unmanned balloons eliminated both the 
danger and the high cost of manned balloon flights. Millikan's 
electroscopes, masterpieces of ruggedness and sensitivity, were 
borne aloft by two balloons; at a certain altitude one of the balloons 
would burst and the other would then bring the equipment gently 
back to earth. During the flight, a simple device continuously 
1 See Appendix A. 


Cosmic Rays 

A radiation from outer space 


recorded the electroscope readings on photographic film, to be 
developed and examined after recovery. 

In the late 1920s and early 1930s the technique of self-record- 
ing electroscopes, carried by balloons into the highest layers of the 
atmosphere or sunk to great depths under water, was brought to an 
unprecedented degree of perfection by the German physicist Erich 
Regener and his group. To these scientists we owe some of the most 
accurate measurements ever made of cosmic-ray ionization as a 
function of altitude and depth (Figs. 1-5 and 1-6). 

Arrowhead Lake 

Fig. 1-4 Principle of Millikan's experiment at Muir Lake and Arrowhead 
Lake in Southern California. An electroscope (at level E) lies 6 feet deeper in 
Muir Lake than in Arrowhead Lake. Since the layer of air between the levels 
of the two lakes weighs as much as 6 feet of water, cosmic rays incident from 
above the level of Muir Lake must traverse the same total mass per unit area 
of absorber (air plus water) before reaching either electroscope. Millikan 
assumed that equal masses per unit area of air or water absorb cosmic rays 
equally. Thus, if no cosmic rays are created in the air layer between the two 
lakes, the number of such rays reaching the two electroscopes should be the 
same. If, however, new rays (dashed lines) were created in the air layer, then 
the electroscope at Arrowhead Lake would record more rays than the one at 
Muir Lake. 










°- icn 






.2 100 





100 200 300 400 500 600 700 800 900 1.000 1,100 

Depth below top of atmosphere, g/cm 

Fig. 1-5 Intensity of cosmic rays as a function of atmospheric depth, as 
measured by Regener and his group with balloon-borne electroscopes. The 
atmospheric depth plotted on the horizontal axis is the mass per unit area of 
the air layer above the electroscope. The vertical scale gives the number of ion 
pairs produced per second by cosmic rays in 1 cm* of air at standard tempera- 
ture and pressure. In these units, the cosmic-ray intensity at sea level is about 2. 


Cosmic Rays 


s 06 








Seo level 


5,000 10,000 

Depth below top of atmosphere, g/cm 



Fig. 1-6 Intensity of cosmic rays under water, as measured by Regencr and 
others. The total mass per unit area of air and water above the electroscope is 
plotted on the horizontal axis. Sea level corresponds to a depth of 1,033 g/cm*. 
The vertical scale gives the number of ion pairs produced per second in 1 cm' of 
air at standard temperature and pressure. 

Atoms and 


Unlikely as it may seem, no serious attempt was made to find out 
what cosmic rays actually are until 16 years after Hess's discovery. 1 
Of course, it would have been difficult to identify the nature of 
cosmic rays with the comparatively crude experimental techniques 
then available. The methods used so successfully in studies of, for 
example, X-rays and radioactivity could not be applied to cosmic 
rays. X-rays and radioactivity came from well-defined sources and 
were subject to experimental control in a laboratory. Cosmic rays, 
on the other hand, seemed to come from everywhere. They could 
not be turned on and off at will as X-rays could be. And, unlike a 
sample of radioactive material, the source of cosmic rays was 
inaccessible, not to say unknown. More important, at ground level, 
where it would have been easiest to study them, the ionizing effects 

1 It was 24 years afterward, in 1936, that Hess shared a Nobel Prize in 
physics for his discovery. 



Cosmic Rays 

of cosmic rays were exceedingly small compared to those caused by 
terrestrial radiations. 

Aside from practical difficulties, however, there was little 
incentive to investigate the problem of the nature of cosmic rays, 
because most scientists believed they already knew the answer. 
What were the reasons for the prevailing misconceptions about 
cosmic rays? And what were the ideas behind the experiments, 
begun in 1929, that eventually led to our present understanding of 
their physical nature? To answer these questions, we must first 
recall what was known in 1929 about atoms and radiations. 


In the 16 years after Mess's discovery, physicists had achieved a 
much clearer understanding of atomic structure. They knew that 
atoms consist of a comparatively heavy, positively charged nucleus 
surrounded by a cloud of negatively charged electrons of almost 
negligible mass. Also, they had discovered that classical physics 
could provide no satisfactory description of the behavior of elec- 
trons in atoms, and they had developed a new theory — known as 
quantum mechanics — to deal with that and other atomic phe- 

As for atomic nuclei, various investigators had measured their 
masses, electric charges, and approximate sizes. Very little was 
known about their structure, however. After the discovery of radio- 
activity, it was clear that at least the nuclei of heavier elements did 
have a structure, that is, consisted of smaller "building blocks." 
For example, occasionally a nucleus of radium disintegrated spon- 
taneously into a nucleus of radon and a nucleus of helium (called 
an alpha particle, a). 

Thus, the radium nucleus was not a "simple" particle. More- 
over, it had been known for 10 years that high-speed o particles 
could break up the nuclei of certain elements such as nitrogen and 
cause them to eject positive particles — protons — identical with 

Atoms and radiations 


hydrogen nuclei. There were speculations that all nuclei were 
formed from the same building blocks. Indeed, since nuclear 
masses were known to be nearly whole-number multiples of the 
proton mass, it was thought that these building blocks might be the 
proton and the (practically massless) electron. 

But scientists did not accept this hypothesis without mis- 
givings, because they were well aware that it presented a number 
of serious difficulties. The puzzle of the constitution of nuclei was 
to be solved a few years later when the British physicist James 
Chadwick discovered the neutron. It then became clear that nuclei 
consist of protons and neutrons; see Appendix G. 


At the time I am speaking of, the word "radiations" referred, as it 
does now, to a variety of different physical phenomena having 
certain features in common. One of the most important character- 
istics of any radiation is its energy. All radiations are energy 
carriers, and they will give up part or all of their energy to an 
appropriate absorber. In many cases the added energy can be 
detected directly in the form of heat, as in the simple example of an 
object exposed to bright sunlight. 

Physicists generally distinguished between two broad groups 
of radiations: corpuscular and electromagnetic. Corpuscular radia- 
tions included beta rays (0), or high-speed electrons emitted in the 
spontaneous decay of certain radioactive atoms; a rays, or helium 
nuclei, emitted in the spontaneous decay of radioactive atoms; and 
artificially accelerated particles such as cathode rays, which are 
electrons emitted by the negative electrode, or cathode, during the 
electric discharge in a vacuum tube. Electromagnetic radiations 
included visible light, infrared and ultraviolet rays, X-rays, and 
gamma rays (7). Gamma rays are among the products of radio- 
active decay, and they are emitted simultaneously with rays 
and a rays. 


Cosmic Rays 

Atoms and radiations 


In earlier years, the main difference between the two groups 
of radiations appeared to be that corpuscular rays were particles 
and electromagnetic rays were waves. By the late 1920s, however, 
the development of quantum physics had brought about a drastic 
revision of this view. Physicists realized that all rays behave in 
some respects as waves and in other respects as individual particles. 
(Particles of electromagnetic radiations had been named photons.) 

The concept of something that behaves both as a wave and as 
a particle is difficult to grasp; fortunately, however, it is not essen- 
tial to our discussion. As it turns out, in fact, cosmic rays contain 
only photons and other particles of very high energy whose wave 
characteristics are never dominant. Thus, in what follows we can 
forget about waves and take the view that corpuscular as well as 
electromagnetic rays consist of individual particles, or quanta. 

Even though, by 1929, the fundamental distinction between 
particles and waves had disappeared, it was still convenient — as 
it is now — to maintain the traditional classification between cor- 
puscular rays and electromagnetic radiations. 

All corpuscular rays are particles to which a definite mass can 
be assigned (9.11 X 10 -28 gram for an electron in a /3-ray beam; 
6.64 X 10 -2 * gram for a helium nucleus in an a-particle beam). 1 
The kinetic energy, or energy of motion, 2 of each particle depends 
on its velocity as well, of course, as on its mass; and the velocities 
vary considerably. Particles of the kind found in many corpuscular 
radiations travel at "relativistic" velocities (that is, at velocities 
comparable to the velocity of light).' According to the laws of 
relativistic, as opposed to classical, mechanics, no material (cor- 
puscular) particle can ever reach the velocity of light (about 
300,000 kilometers/sec in a vacuum), even though it approaches 
this velocity more and more closely as its kinetic energy increases. 

On the other hand, the photons of electromagnetic radiations 
do not possess a finite mass. Although they have different energies, 

1 For powers of 10, see Appendix B. 
1 See Appendix D. 

• an. 

they all travel at the velocity of light. Moreover, although particles 
of finite mass may still continue to exist as particles when they are 
"stopped" by an absorber, photons disappear altogether by, for 
example, transferring their energy to electrons or atomic nuclei. 
AH electromagnetic radiations are alike in another respect: they 
originate from the accelerated motion of electric charges. What 
distinguishes one type of electromagnetic radiation from another 
is simply the different, individual energies of the photons. Photons 
of visible light are more energetic than infrared photons and less 
energetic than ultraviolet photons; these, in turn, are less energetic 
than X-ray photons; and the latter are less energetic than 7-ray 

Of the radiations known prior to the discovery of cosmic rays, 
those emitted by radioactive substances had the highest energies. 
In fact, the energies of individual a particles, particles, and 7-ray 
photons are of the order of millions of electron volts. By contrast, 
photons of visible light have energies of the order of 10 electron 
volts. (One electron volt, abbreviated eV, is the kinetic energy of 
an electron accelerated by a potential difference of one volt. 1 
Cosmic-ray physics involves much greater energies, which are 
measured in millions of electron volts (MeV) or billions of electron 
volts (BeV). 

Ionization by corpuscular rays 

As I have already noted, from the end of the nineteenth century it 
had been the common practice of physicists to detect certain radia- 
tions by their property of ionizing gases. By the late 1920s, the 
manner in which different types of radiation ionize gases was well 
understood. This matter is of fundamental importance to any 
treatment of cosmic-ray phenomena; for until recent years almost 
the only means of studying cosmic radiation was through its 
ionizing effects on gases. Hence, I must ask those of my readers 

1 See Appendix E. 


Cosmic Rays 

Atoms and radiations 


who are already familiar with ionization in gases to bear with me 
through the following discussion. 

Consider first a beam of charged particles (0 rays, say) trav- 
ersing a gas. When one of the particles passes near a molecule of 
the gas, its electric forces disturb the arrangement of the electrons 
in the molecule. If the disturbance is sufficiently violent, one of the 
electrons may break loose and leave behind a positively charged 
molecular fragment, or positive ion. The electron itself may remain 
free for a while and wander about in the gas as a light negative ion, 
or it may attach itself to a neutral molecule and form a heavy 
negative ion. In either case, the charged particle leaves in its wake 
a trail of ions. 

The ion trail can be made visible in the ingenious instrument 
known as the cloud chamber, invented by C. T. R. Wilson in 1911. 
Wilson's original cloud chamber — or expansion chamber, to dis- 
tinguish it from later types — is essentially a glass box with a 
movable wall, or piston (Fig. 2-1). The chamber contains a mixture 
of alcohol vapor and either air or some gas such as argon. If one 
suddenly enlarges the enclosed space by pulling back the piston, 
the gases expand and their temperature drops. If the temperature 
drop is great enough, the alcohol vapor condenses into droplets, 
which form a fog in the chamber. Proper control of the expansion, 
however, will produce a somewhat smaller temperature change, and 

Fig. 2-1 Expansion cloud chamber. 
When the piston is pulled back rapidly, the 
gas and vapor in the chamber expand. The 
resulting drop in temperature is sufficient 
for condensation of the vapor, which, if 
carefully controlled, takes place around 
any ions present in the gas. 

the vapor will condense only around ions. Consequently, a charged 
particle traversing the chamber at about the time of the expansion 
will leave a "fossil track" in the form of visible droplets that 
condense along the particle's trail of ions. 

To be detected, the particle must traverse the chamber at 
some time during the so-called expansion phase. If the particle 
enters the chamber too soon — more than a few milliseconds prior 
to expansion, the ions will diffuse away before the gases are cooled. 
If it enters too late — more than a few milliseconds after expansion, 
the gases will warm up again before the ion trail is formed. Thus, 
the chamber is sensitive for a period of about one-hundredth of a 
second each time it is expanded. 

Fast-moving particles with different electric charges and differ- 
ent velocities produce ion trails of different densities in a gas. For 
a given velocity, the density, or number of ions per unit length in 
the trail, increases with increasing charge. The explanation is fairly 
simple. The electric forces exerted on the electrons of a molecule by a 
particle passing nearby are proportional to the electric charge of the 
particle. Therefore, particles of greater charge cause more violent 
disturbances than particles of smaller charge, and thus they ionize 
a larger number of molecules while traveling a given distance in the 

For a given charge, the ion density decreases with increasing 
velocity. To understand the reason for this effect, consider a particle 
that passes near a molecule. The particle disturbs the molecule 
appreciably only while it is within a certain minimum distance. 
Hence, the time of interaction is longer for a slow-moving particle 
than for a fast-moving one. Obviously, the longer the period of 
interaction, the more effective the force will be in disrupting the 
molecule. Therefore, slower particles ionize more heavily than 
faster particles. 

As I mentioned before, when the kinetic energy of a particle 
increases indefinitely, the velocity of the particle eventually 
approaches that of light. The density of the track left by such a 
particle at first decreases with increasing energy, but then tends to 
a practically constant value characteristic of a particle moving at 

Cosmic Rays 

nearly the velocity of light (Fig. 2-2). Hence, there is a lower limit 
to the amount of ionization produced by any charged particle. The 
ion trail of smallest possible density is one left by a singly charged 
particle (for example, an electron or a proton) moving at nearly 
the velocity of light. Such a particle is called by physicists a 
minimum-ionizing particle. For example, P rays with kinetic ener- 
gies of the order of 1 MeV, which are singly charged and travel at 
about 94 per cent the velocity of light, behave very much like 
minimum-ionizing particles. In air at standard temperature and 
pressure they produce about 50 pairs of positive and negative ions 
per centimeter of path. On the other hand, a particles emitted in 
the radioactive decay of polonium have an energy of 5.3 MeV and 
a velocity only 5.4 per cent that of light; they carry two elementary 
charges and produce about 24,000 ion pairs per centimeter of path 

in air. 1 

Ionization processes similar to those in gases also occur when 
charged particles pass through liquids or solids. But whatever the 
substance — solid, liquid, or gas — the particles lose, in general, a 
very small fraction of their energy in any ionization "event" 
(something of the order of 30 eV in air). Yet the cumulative effects 
of many such events will gradually slow down a particle and 
eventually bring it to rest. In fact, in the late 1920s it was com- 
monly believed that all charged particles traveling through matter 
lost their energy almost entirely by ionization. This belief, as we 
shall see, turned out to be wrong. 

Ionization by photons 

The discussion thus far has concerned only the ionization caused 
by fast-moving charged particles. But what about photons, which 

1 Electrons ejected from a molecule by a moving, charged particle often 
have sufficient energy to produce, in turn, one or several ion pairs. The ioniza- 
tion densities quoted here include both the ion pairs produced directly by the 
"primary" particle and those produced by the "secondary" electrons. Direct 
ionization accounts for something less than half the total. 















1 — 1 — 1 1 1 1 1 1 1 1 1 — 1 1 1 1 1 1 

,...!.. . 



1. . 




1 , 





_ % 




4 2 

a s i. 3 

*• -a 

« GO 

* S 

s a. 


E a 

f. P 


1 "8 



= 8 u -5 


a" 8 


-* .-; 

1 * 


e o. 
.2 < 


2 o 

? > 



4 1 
2 J 

- § 

.2 * 



3 -2 


■a s 



M ° 

f S 

c — 

I § 


° E 


a E 

s ; 

V ■ 

g I 

i s 
£ I 

C >> 

1 g 

? - 

fe .2 

•I s 

2? 3 



5 I 

a .= 

a - 


a 5 

E .5 

8 >, 

* - ,2 

£ £ 

ca so 

a e 

Cosmic Rays 

carry no charge? Photons too can ionize gases, but by a different 
process, a process that occurs much more rarely than the one 
involving charged particles. For example, a 2-MeV electron ionizes 
about 50 molecules for every centimeter of path in air at standard 
temperature and pressure, whereas a photon of the same energy 
travels an average distance of 170 meters before it knocks an elec- 
tron out of a molecule. Furthermore, unlike a charged particle, 
which loses only a minute fraction of its energy in each collision, a 
photon loses most or perhaps all of its energy in a single encounter. 

This basic difference in behavior between photons and charged 
particles is not too surprising, since a charged particle does not 
have to "hit" an electron directly to knock it out of a molecule. 
It is the particle's electrostatic repulsion or attraction acting upon 
the electron that shakes the electron loose, and either of these 
forces is effective at distances much greater than the dimensions 
of the particle itself. The photon, because it lacks an electric 
charge, is not capable of such long-range effects. Nothing happens 
until it makes a "direct hit" on an electron. Because of the small 
size of the target, such an event is naturally rare. But when it 
occurs, it is a much more violent affair than an interaction at a 

In this account of cosmic rays, we shall be concerned for the 
most part with photons whose energy is much greater than the 
minimum necessary to remove an electron from an atom or a 
molecule. It turns out that, with respect to these photons, the 
electrons behave essentially as free particles. In other words, the 
fact that they are part of a complex atomic or molecular structure 
is immaterial. 

The collision of a photon with an electron closely resembles the 
collision of a moving billiard ball with a ball at rest. The photon (the 
moving ball) bounces off at an angle with its energy much reduced, 
while the electron (the ball at rest) is knocked off in a different 
direction (Fig. 2-3). This type of interaction between photons and 
electrons is called the Compton effect, after A. H. Compton, who 
discovered it in 1923. When a beam of high-energy photons trav- 

el loms and radiations 23 

erses a gas, the number of molecules ionized as a result of Compton 
collisions is exceedingly small. The recoil electrons produced in these 
collisions, however, carry off a sizable fraction of the photon 
energy, and they are therefore capable of producing large numbers 
of ions before coming to rest. Thus photons ionize gases indirectly, 
through the few high-speed secondary electrons that they eject 
from gas molecules. 

Absorption and energy loss 

I have been speaking somewhat loosely of the absorption of cor- 
puscular rays and photons by matter. Notice how differently the 
two kinds of rays behave in this respect. 

Charged particles lose their energy bit by bit through a large 
number of ionization events. A beam of particles, all having the 
same mass, charge, and initial kinetic energy, will slow down at 
almost exactly the same rate as they travel through matter. Thus 
they will all come to rest after traversing the same distance. This 
distance, called the range, depends, of course, on the nature and 
initial energy of the particles and on the material of which the 
absorber is made. For the reasons explained in Appendix A, the 
range is often measured in grams per square centimeter rather than 
in centimeters. 

What happens to any given photon that travels through 
matter, however, is largely a question of chance. It may undergo a 

Scattered photon,, S* "~ ~. __ ^ 

Recoil electron 

Fig. 2-3 Compton effect. A photon "collides" with an electron at rest. The 
photon, having given part of its energy to the electron, is scattered; the electron 
recoils in a different direction. The vector sum of the momenta of the scattered 
photon and the recoil electron equals the momentum (see Appendix D) of the 
incident photon. 


Cosmic Rays 

Compton collision almost immediately or after traveling a con- 
siderable distance. Consequently, in a beam of photons that has 
traveled a certain distance, some photons still have the same energy 
they started with, while others have lost a large share of their 
energy in Compton collisions. The latter, moreover, will have been 
deflected away from the beam. In the case of a photon beam, we 
can predict only the average behavior of the particles. We can say 
that a certain fraction of all photons traversing a given mass of 
matter will be lost from the beam. This fraction depends on the 
thickness of the absorber, on the material in the absorber, and, if 
all the photons have the same energy, on the value of that energy. 

If, ideally, the absorber is divided into layers of the same 
thickness, the same fraction of the incident photons is absorbed in 
each layer. Suppose, for example, that the fraction absorbed is 
10 per cent. Starting with a million photons, after the first layer the 
number of photons in the beam will be 900,000 (1,000,000 - 
100,000); after the second layer it will be 810,000 (900,000 - 
90,000), and so on. By plotting the number of surviving photons 
as a function of absorber thickness, one obtains an absorption curve, 
such as that shown in Fig. 2-4. In mathematical language, a curve 
of this general shape is said to be exponential. 

As we have seen, photons of a given energy travel different 
distances before undergoing their first collisions. The average dis- 
tance is called the mean free path, and, like the range of particles, 
it is often measured in grams per square centimeter. One can prove 
(but I shall not attempt to do so here) that the mean free path is 
also equal to the absorber thickness necessary to cut down the 
number of photons to 1/2.7 of its initial value. (The number 2.7 is 
the base of natural logarithms.) Therefore, the shorter the mean 
free path, the steeper the absorption curve. 

Such was the general understanding of the absorption of 
corpuscular rays and photons in 1929. Before turning to the ques- 
tion of cosmic rays and their physical nature, I should like to point 
out that this understanding rested on two important assumptions: 
that charged particles lose energy only by ionization and that 

Atoms and radiations 


photons lose energy only by undergoing Compton collisions. As 
later research was to show, however, at sufficiently high energies 
other absorption processes come into play. High-energy particles 
and photons arc absorbed much more readily than the physicists 
of 1929 supposed. The absorption of high-energy particles and 
photons will be discussed in Chaps. 7 and 10. 



f photons 



Z 400,000 

free parr 


— Mean 












14 IS 

Absorber thickness, number of layers 

Fig. 2-4 Exponential absorption curve (see text). The intersection of the 
curve with the horizontal line that passes through the point at 1 .000,000/2.7 
on the vertical axis determines the mean free path. 

The nature of 
cosmic rays 


The gamma-ray hypothesis 

Most of the early workers in the field of cosmic rays were convinced 
that the rays were simply photons of greater energy than any 
previously discovered. In fact, as late as 1929, the German litera- 
ture commonly referred to cosmic rays as Uliragammaslrahlung 
(ultra-gamma radiation). The reason is easy to understand. Aside 
from cosmic rays, the most penetrating radiation known at the 
time consisted of y rays from radioactive substances. The mean 
free path of y-ray photons in air measured hundreds of meters, 
whereas /3 rays of comparable energies showed ranges of only 
several meters and the ranges of a particles were smaller still. 

Moreover, theoretical computations, based on the assumption 
that high-energy photons were absorbed predominantly through 
Compton collisions, predicted an increasing mean free path with 



Cosmic Rays 

increasing photon energy. In confirmation of this prediction, it had 
been found experimentally that the most energetic y rays were also 
the most penetrating. 1 

Now, absorption measurements in the atmosphere and under 
water by Hess, Kohlhorster, Millikan, Regener, and others had 
shown the cosmic radiation to be more penetrating than any other 
(see Figs. 1-5 and 1-6). It was thus natural to think that cosmic 
rays were of greater energy than the most energetic photons found 
among y rays. Some physicists, willing to trust the theory of the 
Compton effect in a range of energies where it had not yet been 
tested experimentally, estimated the energies of the hypothetical 
cosmic-ray photons from the shape of their absorption curve. 
Cosmic rays, they concluded, were a mixture of photons with 
energies ranging from 20 or 30 to several hundred MeV. This esti- 
mate was the basis of a most provocative suggestion put forward by 
Millikan in 1928. According to Millikan, cosmic rays were born of 
the energy released during the synthesis of heavier elements from 
primordial hydrogen spread throughout the universe. 

Millikan's hypothesis 

As explained in Chap. 2. a beam of high-energy photons, all having 
the same energy and all traveling in the same direction through 
matter, are absorbed exponentially (Fig. 2-4). The absorption 
curve is characterized by a mean free path whose length, for a 
given absorber material, depends on the photon energy. But even 
if the cosmic-ray photons all carried the same energy (which is very 
unlikely), their absorption could not be represented by a simple 
exponential curve, because cosmic rays do not form a parallel 
beam, but instead fall on the atmosphere from all directions. It is 
not difficult to correct for this effect and to compute the theoretical 

1 The theory of the Compton effect was published in 1929 by Oscar Klein 
of Sweden and Yoshio Nishina of Japan. A preliminary and less accurate 
theory had been developed by P.A.M. Dirac of Kngland in 1926. 

The nature of cosmic rays 

absorption curves of photons incident from all directions and 
characterized by a particular mean free path. 

Millikan noted that the actual absorption curve of cosmic rays 
did not correspond to any single curve thus computed. However, 
it could be represented by the sum of the absorption curves of three 
groups of photons with mean free paths of 300, 1,250, and 2,500 
g/cm 2 . Using Dirac's theory of the Compton effect (Klein and 
Nishina had not yet published theirs), Millikan arrived at energies 
of about 26, 110, and 220 MeV for the three respective groups of 
photons. He therefore concluded that cosmic radiation was for the 
most part a mixture of photons with those energies. 

While searching for clues to the origin of the photons, which 
appeared to come in equal numbers from every region of the sky, 
Millikan was led to the following speculation: Interstellar space is 
filled with very dilute hydrogen gas. Conceivably, out of this gas 
the atoms of the heavier elements might continuously evolve by a 
spontaneous process of fusion. Once in a while, for example, four 
hydrogen atoms might meet and fuse to form a helium atom. 

A helium atom weighs slightly less (about 4.8 X 10 -26 gram 
less) than four hydrogen atoms. According to Einstein's principle 
of equivalence between mass and energy. Ibis means that if a 
helium atom forms from four hydrogen atoms, an amount of energy 
equal to 4.8 X 10~ 28 times the square of the velocity of light is 
released. 1 In the usual units, the amount of released energy is 
27 MeV. Now, allowing for experimental error, 27 MeV was just 
the energy of the first of the three groups of photons found in 
cosmic radiation. Consequently, Millikan concluded that these 
photons result from the synthesis of helium in interstellar space. 

Thus encouraged, Millikan explored the possibility of finding 
a similar explanation for the origin of the other two groups, and 
he was remarkably successful. Among the most abundant elements 
in the universe are nitrogen and oxygen. One atom of nitrogen 
weighs about 1.8 X 10 _2S gram less than fourteen atoms of hydro- 
gen. From this mass defect one can calculate that the fusion of 

1 See Appendix D. 


Cosmic Hays 

fourteen hydrogen atoms into one atom of nitrogen would produce 
about 100 MeV of energy. Similarly, the synthesis of one oxygen 
atom from sixteen atoms of hydrogen would release an energy of 
about 120 MeV. 

These values are close to each other and to the estimate for 
the photons of Millikan's second group. Quite naturally, Millikan 
concluded that these photons also resulted from the synthesis of 
heavier elements, namely, nitrogen and oxygen. But that was not 
all. For, as other calculations showed, if twenty-eight hydrogen 
atoms fused to form one atom of silicon, another abundant element, 
the process would release an amount of energy very nearly equal 
to the energy of the photons in the third group. This set of coin- 
cidences, which turned out to be purely accidental, convinced 
Millikan that cosmic rays were indeed the "birth cry" of atoms 
being continuously created in space. 

The coincidence experiments of Bothe and Kohlhorster 

Of course, from time to time physicists expressed doubts about the 
nature of cosmic rays. But the prevailing view of cosmic rays as 
high-energy photons was not seriously challenged until 1929, when 
the German physicists Walther Bothe and W. Kohlhorster pub- 
lished a paper titled "Das Wesen der Hoehenstrahlung" ("The 
Nature of the Badiation from Above"). In this paper, Bothe and 
Kohlhorster related certain experiments and arrived at certain 
conclusions that marked a turning point in the history of cosmic- 
ray research. The experiments were made possible by an important 
technical development earlier the same year: the invention of a 
convenient instrument to count cosmic rays. 

Electroscopes of the type used in earlier work on cosmic 
rays could detect only the combined ionizing effects of large num- 
bers of particles over relatively long periods of time. They could 
not detect individual particles, because the passage of a single 
particle through an electroscope does not produce enough ioniza- 

The nature of cosmic rays 


tion to cause an observable movement of the electroscope leaves. 
There were, of course, instruments capable of detecting individual 
ionizing particles. For many years, physicists had been using the 
Wilson cloud chamber, which I have already described, and had 
been "counting" a particles by observing the tiny sparks, or 
scintillations, produced when the particles hit a fluorescent screen. 
Moreover, Hans Geiger of Germany, while working in the 
laboratory of Ernest Rutherford at the University of Manchester 
from 1906 to 1912, had collaborated in the development of elec- 
trical counting devices known as point counters. A point counter is 
basically a thin, pointed rod projecting into a metal box (Fig. 3-1). 
A battery or other voltage source maintains the rod at a positive 

Fig. 3-1 Particle counter of the type used in early experiments on radio 
activity. It consists of a metal box containing a pointed rod that is held in 
place by an insulator. Ionizing particles enter the box through a window in 
front of the point. The window is covered by a thin foil so that the counter can 
be operated with different gases and at different pressures. The case is con- 
nected to the negative terminal of a high-voltage battery with its positive 
terminal grounded. The rod, connected to an electroscope, is grounded through 
a large resistor; thus the rod is at a positive voltage with respect to the case. 
When a discharge occurs in the counter, the electroscope wires undergo a 
sudden deflection. The wires then quickly return to their original position as 
the charge leaks to ground through the resistor. 


Cosmic Rays 

electric potential with respect to the box. When the potential 
difference is less than a certain critical value (usually of the order 
of 1,000 volts, depending on the design of the counter and on the 
gas pressure), the electric field simply sweeps positive ions toward 
the negatively charged walls and negative ions (electrons) toward 
the positively charged point, just as ions of opposite sign are 
swept toward the case and the leaves of a charged electroscope. 

The drift of the ions, toward the walls or the point, is com- 
paratively slow because both positive and negative ions collide 
frequently with the gas molecules. But at potential differences 
above the critical value something new happens in the vicinity of 
the point, where the electric field is strongest. In the short time 
interval between collisions, some of the electrons acquire sufficient 
kinetic energy to break up the gas molecules with which they 
collide. In these collisions more electrons are ejected; they are 
accelerated toward the point and become capable of ionizing the 
gas molecules they encounter. 

The result is a growing avalanche of ions. The gas suddenly 
acquires a very large electrical conductivity, and a discharge takes 
place. The value of the point counter lies in this multiplication 
effect, whereby the few ions produced by the passage of a single 
ionizing particle suffice to trigger a discharge that can easily be 
detected by, for example, a Wulf electroscope connected as shown 
in Fig. 3-1. 

However, none of the instruments that I have mentioned was 
well suited to cosmic-ray studies. The cloud chamber was difficult 
to run and was sensitive for no more than a small fraction of a 
second at each expansion. The fluorescent material of scintillation 
counters detected only those particles capable of ionizing much 
more heavily than most of the particles associated with 'the cosmic 
radiation. The point counter was not very stable and could not be 
made sufficiently large to be of much use in the study of a radiation 
whose intensity was as small as that of cosmic rays. 

The answer to the needs of cosmic-ray physicists came in 1929 
with the invention, by Geiger and one of his students, W. Muller, 

The nature of cosmic rays 


of the so-called Geiger-Muller counter (or G-M counter for short). 
This instrument, one of the most widely used tools of experimental 
physics in the twentieth century, essentially consists of a metal 
tube with a thin metal wire stretched along its axis (Fig. 3-2). The 
tube is first evacuated and then filled with a gas at a pressure of 
about one-tenth of an atmosphere. Under operating conditions, the 
wire is held at a positive potential with respect to the tube. 

The G-M counter works on the same principle as the point 
counter; that is, it makes use of the cascade effect, in which a single 
ion pair produced in the gas triggers a discharge. The operating 
voltage is usually between 1,000 and 1,500 volts, and the counter 
is sensitive over practically its whole volume. The G-M counter is 
comparatively easy to build, is much more reliable than the point 
counter, and can be made in a variety of sizes, up to several inches 
in diameter and several feet in length. 

In their experiments, Bothe and Kohlhorster had set up two 
G-M counters, each connected to an electroscope, in order to 
observe cosmic rays. They noticed that the counters, when placed 
one above the other a small distance apart (Fig. 3-3), often dis- 





Fig. 3-2 The Geiger-Muller counter: metal tube 7V glass insulators C; thin 
wire W; tube for evacuating and filling the tube E. Electrical connections are 
similar to those shown in Fig. 3-1. 


Cosmic Rays 





Electroscope 1 


Electroscope 2 


Fig. 3-3 Two G-M counters placed one above the other and connected to two 
electroscopes. Simultaneous deflections of the electroscope leaves indicate 
simultaneous discharges of the two counters, or coincidences. 

charged simultaneously. These simultaneous discharges, or coin- 
cidences, could not be due to chance, or at least not entirely to 
chance, because they became much less frequent when the distance 
between the two counters was increased. 

A photon could, in principle, produce a coincidence by a double 
Compton effect (Fig. 3-4a). However, the probability of a Compton 
collision in the wall or in the gas of a counter is very small, and the 
probability of two Compton collisions is quite negligible. As Bothe 
and Kohlhorster rightly concluded, the observed coincidences 
must be due to the passage of individual charged particles through 
both counters (Fig. 3-46). Furthermore, the particles could not be 
ordinary a or /S rays, because the counter walls (1 mm thick and 
made of zinc) would stop all such particles. 

In itself, this result did not contradict the view that the 
primary radiation, the cosmic radiation falling on the atmosphere 
from outer space, consisted of high-energy photons. Since photons 
undergo Compton collisions in the atmosphere, the observed 
ionizing particles might have been the recoil electrons arising from 
such collisions. As I mentioned before, the energies of the hypo- 

The nature of cosmic rays 


thetical primary photons were believed to range from 20 or 30 to 
several hundred MeV. The recoil electrons of these photons would 
have more than enough energy to traverse the counter walls. 

To check upon this possibility, Bothe and Kohlhorster placed 
a gold block 4.1 cm thick between the counters. (Gold was chosen 
because of its high density and correspondingly high stopping 
power.) Now only particles with a range greater than 4.1 cm of 
gold could traverse both counters and produce coincidences. The 
experimenters found that the rate of coincidences was still 76 per 
cent of what it had been without the block. In other words, 76 
per cent of the charged particles present in the cosmic radiation 
near sea level could penetrate 4.1 cm of gold. 

(a) (b) 

Fig. 3-4 Photon and charged-particlc coincidences, (a) A photon produces a 
Compton electron in the upper counter and then another Compton electron 
m the lower counter. Typically, the probability of each Compton collision is of 
the order of 5 per cent. This means the number of pulses (recording Compton 
electrons) in each counter is about one-twentieth the number of photons that 
traverse the counter; the number of coincidences is therefore approximately 
"•25 per cent (1/20 X 1/20 = 1/400) of the number of photons that traverse 
both counters. (6) Every particle passing through both counters produces a 


Cosmic Rays 

This result was very surprising; for according to even the most 
generous estimates, only a small fraction of the recoil electrons 
found at any point in the atmosphere should have had a greater 
range. Bothe and Kohlhorster concluded that the then-current 
ideas concerning the nature of cosmic rays were probably wrong 
and that the primary cosmic radiation itself consisted oj charged 
particles rather than photons. 

Photons versus charged particles 

Consider for a moment the consequences of the two contrasting 
hypotheses, namely, that primary cosmic rays are (1) photons, the 
generally accepted view, or (2) charged particles, as proposed by 
Bothe and Kohlhorster. Remember also that, in 1929, high-energy 
photons were thought to be absorbed through Compton collisions 
alone and charged particles supposedly lost energy only through the 
ionization of matter. (For the sake of simplicity, the rays may be 
assumed to travel vertically downward, instead of coming from all 

1. Near sea level the absorption curve of cosmic radiation 
resembles the absorption curve of photons having a mean free path 
of about 300 g/cm 2 in air or water. Hence one can simplify matters 
by assuming that cosmic-ray photons found near sea level have a 
single energy corresponding to this mean free path. Calculations 
based on the theory of the Compton effect 1 predict an energy of 
approximately 60 MeV for photons with a mean free path of 300 
g/cm 2 in air or water. As a further simplification, assume that when 
a Compton collision occurs, the photon disappears by giving up all 
its energy to an electron. Then each recoil electron will have an 
energy of 60 MeV and a corresponding computed range of 30 g/cm 2 
in air or water. 

1 That is, the theory developed by Klein and Nishina. It is important to 
keep this distinction in mind when comparing the energy estimates quoted 
here and those used by Millikan (page 29). 

The nature of cosmic rays 


At any point in the atmosphere (or other absorber) the radia- 
tion is, then, a mixture of photons and electrons. This is shown 
graphically in Fig. 3-5a, where dashed lines represent photons and 
solid lines represent recoil electrons. (The figure assumes the recoil 
electrons to be traveling in the same direction as the photons from 
which they received their energy.) 

Now suppose the horizontal line at represents sea level ; the 
line at 100, a depth of 100 g/cm 2 (about 333 feet) below sea level; 
and the line at 200, a depth of 200 g/cm 2 (666 feet) below sea level. 
How do the so-called intensities of cosmic radiation at the three 
levels compare with one another? Since there are six recoil electrons 
at sea level and four at 100 g/cm s and three at 200 g/cm 2 below sea 
level, the intensities at the three levels (as measured with an 
electroscope or a single counter) are in the ratio of 6:4:3. These 
are approximate numbers because the figure shows only 60 photons, 
which is much too small a sample to serve as a basis for accurate 

Using a more adequate sample — 6,000 photons, say — the 
ratio would be 600:430:336. These values correspond to three 
points on the absorption curve shown in Fig. 3-6. Note that even 
though this is the "absorption curve of cosmic rays," it docs not 
actually give the number of primary photons found at the different 
levels; instead, it gives the number of their recoil electrons. The 
curve does not include photons, of course, because the discharge 
rate of an electroscope, as well as the counting rate of a single 
G-M counter, depends only on the number of charged particles that 
pass through the instrument. 

2. If the primary cosmic radiation consists of charged par- 
ticles, these particles must have different initial energies; for if 
they all had the same energy, they would also have nearly identical 
ranges and would all stop at the same level. In Fig. 3-56 the energy 
distribution has been chosen in such a way that the number of 
particles present at a given level is the same as the number of recoil 
electrons at the same level in Fig. 3-5a. For this particular energy 
distribution, the response of an electroscope or of a single G-M 

The nature of cosmic rays 

600 1 


1 ' 

1 ' 



8 inn 









100 200 300 

Depth below sea level, g/cm 2 

Fig. 3-6 Number of charged particles present at various depths below sea 
level. As explained in the text and illustrated in Fig. 3-5, this number is the 
same in the case of a radiation consisting of photons and in the case of a radia- 
tion consisting of charged particles with an appropriate energy distribution. 

counter at any given level is the same as in case 1. In other words, 
the absorption curve of the corpuscular radiation, as usually 
measured, is identical to that of the photon radiation (Fig. 3-6). 

Although the hypothesis of a primary photon radiation and 
the hypothesis of a primary corpuscular radiation lead to entirely 
different pictures of the radiation after it has passed through a 
layer of matter (Fig. 3-5a and 6), the absorption measurements 
taken before 1929 provided no way of distinguishing between the 
two hypotheses. 

The significance of the coincidence experiments of Bothe and 
Kohlhorster was that they afforded, for the first time, the possi- 
bility of distinguishing between the consequences of the opposing 


Cosmic Rays 

hypotheses. If all primary cosmic rays are photons of 60 MeV 
energy, no coincidences whatever would have been recorded with 
4.1 cm of gold between the counters; 60-MeV electrons cannot 
travel that far through gold. But if primary cosmic rays are charged 
particles, then in order to penetrate as far as they do through the 
atmosphere and water, many of them must have energies much 
greater than 60 MeV. Consequently, they could also have pene- 
trated 4.1 cm of gold and thereby have caused the observed 

It is true that the proponents of the photon hypothesis thought 
cosmic rays were a mixture of photons of different energies, some 
greater than 60 MeV. A small fraction of the recoil electrons, 
therefore, would have passed through the gold block used in the 
experiments. Although this circumstance makes the interpretation 
of the experiments a little less direct, it does not change the conclu- 
sion. The crucial point is tins: An absorber that produces only a 
minor change in the radiation intensity when placed above an 
electroscope or a single G-M counter produces completely differ- 
ent effects — depending on whether primary cosmic rays are 
photons or charged particles — when placed between two counters. 
If the rays are photons, the absorber drastically reduces the coin- 
cidence rate (even though some of the Compton electrons may have 
enough energy to traverse the absorber). If the rays are charged 
particles, the absorber affects the coincidence rate only slightly. 

The reader can satisfy himself on this point by going back to 
Fig. 3-5. Suppose first that two thin-walled counters lie one above 
the other at sea level with no absorber between them. In Fig. 3-5a 
and b one finds at sea level six ionizing particles capable of 
traversing both counters and thus producing coincidences. Now 
consider what happens when an absorber (equivalent, say, to 25 
g/cm 2 of water) is placed between the counters. Suppose first that 
primary cosmic rays are charged particles. Five of the six particles 
present at sea level reach level A, 25 g/cm 2 below sea level (Fig. 
3-56). This means the number of coincidences has decreased by 

The nature of cosmic rays 


Suppose next that primary cosmic rays are photons. Only one 
of the six recoil electrons present at sea level reaches level A 
(Fig. 3-5o). In other words, only one electron traverses the 
absorber. Since neither the electrons stopped by the absorber nor 
the Compton electrons generated in it can produce coincidences, 
the absorber must cut down the number of coincidences by five- 

Although this discussion is based on oversimplified assump- 
tions and on a very small statistical sample, it is certainly true that 
25 g/cm 2 of water placed between the counters should produce a 
large decrease in the coincidence rate if primary cosmic rays are 
photons and hardly any decrease at all if primary cosmic rays are 
charged particles. 

There was one serious objection to the conclusions reached by 
Bothe and Kohlhorster. The interpretation of their experiments 
was based on an arbitrary extrapolation of the known properties of 
photons and electrons at low energies. Bothe and Kohlhorster 
themselves were well aware of the potential pitfalls involved. It 
was conceivable, for example, that the energies of cosmic-ray 
photons might be much greater than those computed from their 
mean free path according to the equation of Klein and Nishina, 
which was known to be valid for energies of the order of 1 MeV. If 
such were the case, the secondary electrons would have had a 
greater range, more of them would have penetrated the gold block 
between the counters, and they might have produced much the 
same coincidence effects as a primary corpuscular radiation. 

For this reason it would be misleading to claim that the 
experiment of Bothe and Kohlhorster had proved the corpuscular 
nature of cosmic rays. In fact, the two hypothetical pictures of 
cosmic rays I have sketched here turned out to be almost equally 
far from reality. Cosmic rays are indeed charged particles; but they 
behave very differently than charged particles were thought to 
behave at the time. 

Nevertheless, the work of Bothe and Kohlhorster was a mile- 
stone in the history of cosmic rays. The particular interpretation 


Cosmic Rays 

they gave to the experimental results was not the important thing. 
What mattered most was that for the first time physicists had 
attempted to determine the nature of cosmic rays experimentally. 
The results of the experiments had failed to provide any support 
for the accepted view of cosmic rays as high-energy photons and 
had thus thrown the question wide open. 

Puzzling clues 


The history of cosmic rays in the following years is so closely inter- 
twined with my own life as a scientist that I can hardly escape the 
temptation to insert some personal notes in this account. 

When I happened to read the paper by Bothe and Kohlhorster, 
I was 24 years old and had just completed a year as an assistant to 
Prof. Antonio Garbasso, director of the physics laboratory of the 
University of Florence, in Arcetri, Italy. Among my colleagues 
were Gilberto Bernardini and Giuseppe Occhialini, of whom more 
will be heard later in this story. None of us was yet committed to 
any long-range research program. The paper of Bothe and Kohl- 
horster came like a flash of light revealing the existence of an unsus- 
pected world, full of mysteries, which no one had yet begun to 
explore. It soon became my overwhelming ambition to participate 
in the exploration. 

Within a few months I had built my first G-M counters, had 
devised a new method for recording coincidences (based on the use 
of vacuum tubes; see Fig. 4-1), and had begun some experiments. 
The summer of 1930 took me to Charlottenburg, Germany, for 



Cosmic Rays 

several months of work in Bothe's laboratory at the Physikalische- 
Technischc Reichsanstalt. There I repeated, with greater precision 
and some improvements, Bothe and Kohlhorster's original experi- 
ment. The improvement consisted essentially in making a direct 
comparison of the coincidence rates recorded when an absorber 

Fig. 4-1 Vacuum-tube coincidence circuit greatly reduces the number of 
chance coincidences recorded by G-M counters (see text). Under operating 
conditions, current flows from the positive terminal of the battery B through 
the resistor H and three tubes T,, It, Ttloa ground. This current produces a 
large voltage drop across the resistor, and at point A the potential is nearly 
that of the ground. When one of the G-M counters, d, say, is discharged, the 
grid of the corresponding tube 7\ becomes temporarily negative and the tube 
no longer conducts current, which, however, continues to flow through T» 
and IV If the resistor is sufficiently large, the current changes only slightly as a 
result. The simultaneous discharge of two G-M counters also fails to produce 
a significant change in the current. But if all three counters discharge simul- 
taneously, then all three tubes become nonconducting, the current through R 
stops altogether, and the potential at A suddenly becomes equal to that of the 
positive battery terminal. This voltage jump at A thus signals the occurrence 
of a threefold coincidence; it can be detected by a voltmeter, or it can be 
made to operate a mechanical recording device. The circuit can be adapted to 
record coincidences involving any number of counters. 

Puzzling clues 


was above the counters with those recorded when the same absorber 
was between the counters. 

Upon returning to Arcetri, I continued to experiment along 
similar lines. About these early experiments I shall say only that 
they strengthened my feeling of wonder, almost of awe, before the 
dimly perceived new facts that were beginning to emerge. I was 
convinced that cosmic rays would turn out to be something funda- 
mentally different from the other known radiations, unless the 
properties of photons and charged particles changed drastically as 
their energies increased. This conviction was the motivation for 
two experiments I should like to describe here. 

Cosmic-ray penetration of lead 

Bothe and Kohlhorster had shown that a large fraction of the 
cosmic-ray particles found near sea level were capable of traversing 
4.1 cm of gold, and my own work in Bothe's laboratory had con- 
firmed their results. Curious to find out whether, perchance, a few 
of these particles might have enough energy to traverse much larger 
thicknesses of matter, say, 1 meter of lead, I arranged three G-M 
counters along a vertical line (Fig. 4-2) . The total thickness of lead 
between the counters was initially 25 cm. By piling up a number 
of lead bricks, I could increase the absorber thickness to a total 
of 1 meter, and I found that 60 per cent of the cosmic-ray particles 
capable of traversing 25 cm of lead also traversed 1 meter of lead. 
Since 25 cm of lead absorbed less than half the radiation, not only 
a few (as I had thought) but a very sizable fraction of the cosmic- 
ray particles found near sea level had a range greater than 1 meter 
of lead. 

Considering that the maximum range in lead of rays from 
radioactive substances is only a fraction of a millimeter, one can 
easily appreciate the surprise caused by this result. In the first place, 
cosmic-ray particles, whatever their nature, had to be many orders 
of magnitude more energetic than /3 rays. In fact, they had to 


Cosmic Rays 

possess energies in the range of billions of electron volts. These 
energies were greater than any that could possibly be released in 
the synthesis of elements. And for many physicists it was a great 
disappointment when, in the face of the experimental evidence, 
they were forced to abandon Millikan's fascinating theory about 
the origin of cosmic rays. Moreover, the 7-ray hypothesis, already 
shaken by the work of Bothe and Kohlhorster, had received an 
even more serious blow. 

Incidentally, some readers may wonder why three counters 
rather than two were employed in the experiment I have just 
described. The purpose of the third counter was to reduce the 
number of unwanted, chance coincidences, that is, coincidences 
caused by different particles crossing the counters almost simul- 
taneously. An electronic coincidence device, Fig. 4-1, made it 
possible to cut down the number of chance coincidences con- 
siderably. Even so, with two counters more than 1 meter apart, the 
number of such events would have been greater than the number of 
true coincidences resulting from the passage of single cosmic-ray 





l m L *° d 


Fig. 4-2 Experimental arrangement used to demon- 
strate penetrating power of cosmic-ray particles. The 
number of lead bricks between counters can be varied to 
form an absorber as much as 1 meter thick. Only charged 
particles capable of traversing the absorber produce coin- 


Puzzling clues 


particles through both counters. With a third counter in between, 
the number of true coincidences remained unchanged, but the 
number of chance coincidences was drastically reduced. 

Secondary effects of cosmic rays 

The second experiment followed from observations that seemed to 
suggest that high-energy cosmic-ray particles occasionally pro- 
duced secondary ionizing particles in matter. Perhaps the most 
significant observations were those made in 1929 by the Russian 
physicist D. Skobeltzyn. Working with a cloud chamber placed in 
a magnetic field, Skobeltzyn had photographed the tracks of un- 
usually energetic negative particles passing through the chamber. 
The curvatures of the tracks (Chap. 5) indicated particle energies 
much greater than those of ordinary rays. Skobeltzyn suggested 
that the tracks were probably left by electrons recoiling from 
Compton collisions with the hypothetical cosmic-ray photons. 

Skobeltzyn also pointed out the occasional appearance, in his 
pictures of high-energy particles, of two and in one case, three 
tracks in the same picture. It was possible to explain the multiple 
tracks by assuming that a recoil electron had undergone one or 
more collisions somewhere near the cloud chamber and in these 
collisions had ejected secondary particles of sufficient energy to 
penetrate the chamber wall. 

To investigate the suspected production of secondary particles 
by cosmic rays in matter, I placed three G-M counters in a tri- 
angular array (Fig. 4-3). The three counters could not be dis- 
charged by a single particle traveling in a straight line, and yet, 
when completely surrounded by lead, the array recorded a large 
number of coincidences (some thirty-five per hour). With the upper 
part of the lead shielding removed, the coincidence rate fell almost 
to zero. The coincidences could only have been the result of two or 
more ionizing particles emerging simultaneously from the lead, a 
direct proof that cosmic radiation did give rise to secondary ion- 


Cosmic Rays 

izing rays. But even with the top lead removed, there was a small 
residual effect of roughly two coincidences per hour, which at the 
time I could not fully explain. It later turned out that this effect 
was due to the production of secondary rays in the atmosphere 
(see Chap. 12). 

The unexpected finding was the great abundance of secondary 
rays, as evidenced by the high coincidence rate. In order to appre- 
ciate why this was a cause for surprise, one must recall the novelty 
of the result. The only known interaction of high-energy photons 
was the Compton effect. A Compton collision produces a single 
recoil electron, and therefore it could not account for the groups of 
particles found coming out of the lead absorber. The occurrence 
of two Compton collisions one next to the other and involving the 
same photon is such a rare phenomenon that it can be ignored 
altogether. Charged particles do knock electrons out of atoms — 
this, in fact, is the common ionization process; but these electrons 
ordinarily have a very small energy. Only on rare occasions is an 
electron ejected with sufficient energy to escape from the absorber. 

Thus, nothing of what was known at the time could explain 
the abundant production of secondary particles revealed by the 
experiment. Actually, the results of this experiment appeared so 
incredible to the editors of the scientific journal to which I had 

Fig. 4-3 Triangular array of G-M 
counters used in the first experiment 
demonstrating the production of sec- 
ondary particles by cosmic rays. At least 
two charged particles emerging simulta- 
neously from the lead are needed to 
produce a coincidence. One of them 
may be a primary particle, but the other 
must have been produced in the lead. (If 
the upper section of the lead shielding is 
removed, the coincidence rate falls 
nearly to zero.) 

Puzzling clues 


first submitted my paper that they refused to publish it. The paper 
was later accepted by another journal. 

Cosmic-ray showers 

But this was just the beginning, for it soon became clear that the 
interactions of cosmic radiation with matter give rise to effects far 
more complex and unusual than anyone had the right to infer from 
my counter experiments. The discovery of these complexities was 
the first result of a new experimental device that was to play a most 
important role in the history of cosmic rays: the counter-controlled 
cloud chamber. 

The cloud chamber (see page 18), because it shows the tracks 
of charged particles, is really the ideal tool for finding out in detail 
what occurs in high-energy interactions. When applied to cosmic- 
ray phenomena, however, it has a serious drawback. The chamber 
must expand in order to become sensitive to the passage of an 
ionizing particle. 1 As I mentioned earlier, at each expansion the 
chamber is sensitive for approximately only 0.01 second. Cosmic- 
ray particles, on the other hand, arrive at the comparatively slow 
rate of about one per minute per square centimeter at sea level. 
Consequently, the probability of catching a cosmic-ray particle 
during the expansion phase (see page 19) is very small. Even 
smaller is the probability that this particle will do something of 
interest where the cloud chamber can "see" the results of the 

Obviously, the yield of useful pictures would be increased 
enormously if the chamber expanded at the right time — imme- 
diately after a particle had passed through. It turns out that this 
timely expansion can be brought about by means of the coincidence 
technique. One can, for example, place single G-M counters above 

1 Some years after, in the late 1930s, cloud chambers capable of continuous 
operation (diffusion chambers) were developed. They had, however, a very 
limited application to cosmic-ray research. 


Cosmic Rays 

and below the chamber, connect both counters to a coincidence 
circuit, and use the output signal of the circuit to operate a fast 
release mechanism (Fig. 4-4). More easily said than done, of course! 
The first counter-controlled chamber represented a major tech- 
nical achievement. It was constructed in 1932 and 1933 at the 
Cavendish Laboratory of Cambridge University by P. M. S. 
Blackett and Giuseppe Occhialini. Blackett was already well 
known for his cloud-chamber studies of radioactivity. Occhialini, 
who had become familiar with the technique of G-M counters 
while working in Arcetri, was just beginning his scientific career. 
The 1933 paper in which Blackett and Occhialini described 
their first observations with the counter-controlled chamber 
marked another milestone in the history of cosmic-ray research. I 
shall come back to it in a later chapter. Here I wish to mention 
only one result. A number of pictures showed the tracks of many 
particles that clearly resulted from the interaction of a single high- 
energy cosmic ray somewhere in the vicinity of the chamber 
(Fig. 4-5). These groups of particles, or showers, were unquestion- 
ably the cause of the coincidences between counters out of line that 
I had observed previously. 

Fig. 4-4 Counter-controlled 
cloud chamber. A cosmic-ray par- 
ticle passing through the two G-M 
counters < .', and G 9 and through the 
cloud chamber C produces a coin- 
cidence. The signal pulse from the 
coincidence circuit promptly trig- 
gers the expansion of the chamber, 
before the ion pairs left by the 
particle have time to diffuse away. 
Vapor condensation around the 
ions produces a visible trail of 
droplets along the track of the 

Puzzling clues 


Fig. 4-5 Photograph obtained by Blackett and Occhialini with their counter- 
controlled cloud chamber. The chamber is situated between the poles of an 
electromagnet. Sixteen separate tracks of secondary particles enter the chamber 
simultaneously; they originate above the chamber, being produced, apparently, 
in the copper coils of the magnet (not shown in the picture). The curvature of 
the tracks is caused by the magnetic field; the tracks of positive particles 
curve to the right, the tracks of negative particles to the left. (From P. M. S. 
Blackett and Giuseppe Occhialini, Proceedings of the Royal Society (London), 
vol. A139. p. 699, 1933.] 

52 Cosmic Rays 

A summary 

Let me summarize briefly what had been learned thus far: 

1. The 7-ray hypothesis had no experimental foundation. 

2. Much of the cosmic radiation found near sea level consisted 
of charged particles that possessed energies of 1 BeV or more and 
were capable of traversing very large thicknesses of matter. 

3. Cosmic rays frequently gave rise to very complex processes 
in which large numbers of secondary particles were generated. 

The abundant production of secondary rays meant that a 
significant fraction of the particles observed near sea level had been 
produced in the atmosphere. At the time, many of us were inclined 
to consider the most penetrating particles (those capable of travel- 
ing large distances in lead) to be the real primary cosmic rays and 
the "softer" particles (those absorbed by a few centimeters of lead) 
to be secondary in origin. However, we clearly realized the possi- 
bility that only a few — and perhaps none — of the primary par- 
ticles might be able to penetrate any great distance through the 
atmosphere without undergoing collisions. In that case practically 
all the particles found near sea level would be secondary products of 

The problem of the nature of cosmic rays, which was becoming 
much more complicated than anyone had imagined it could, really 
involved two different questions: 

1. What was the nature of the primary radiation falling on the 
atmosphere from outer space? 

2. What was the composition of the local radiation observed 
in the atmosphere? 

Even if all primary rays were of the same kind, it was clear 
that the local radiation, much of which was secondary in origin, 
was likely to be quite complex. Indeed, in the next two decades it 
was to be found that the local radiation contained not only all the 
radiations previously known, and several kinds of rays that theo- 
rists were to dream about, but many others as well. 

Using the earth 

as a magnet to 

analyze cosmic rays 


Cosmic-ray research had begun with the balloon flights of Hess and 
Kohlhorster. As the years went by, physicists studied cosmic rays 
at higher and higher altitudes. Sounding balloons, as we have seen, 
replaced manned balloons; later, sounding rockets were used. In 
recent years, artificial satellites and space probes have taken coun- 
ters and other recording devices clear of the atmosphere and 
sometimes millions of miles into interplanetary space. But by the 
time physicists had arrived at the stage where their instruments 
were making direct observations of primary cosmic rays, they 
already knew a great deal about these rays. They had obtained 
much of their information by means of the earth's magnetic field. 
The magnetic field surrounding the earth resembles the field of 
a magnetic dipole — the field that would be produced by a bar 
magnet short in comparison with the radius of the earth — located 



Cosmic Rays 

near the earth's center and with its north pole toward the geo- 
graphic south and its south pole toward the geographic north. 
More precisely, the center of the terrestrial magnetic dipole is dis- 
placed about 200 miles from the center of the earth, and its axis 
lies at an angle of about 11° from the geographic axis (Fig. 5-1.) 
The field itself extends far beyond the reaches of the atmosphere. 
At an altitude of 1,000 miles its strength is still about one-half the 
field strength at the earth's surface. By contrast, the remaining air 
mass above 20 miles constitutes less than 1 per cent of the total 

Thus, the primary cosmic rays encounter the magnetic field 
well before they have a chance to collide with the molecules of the 

Axis of 

Geomagnetic axis 

Geomagnetic equator 

Geographic equator 

Fig. 5-1 The magnetic field of the earth resembles the field that would be 
produced by a relatively short bar magnet (or dipole) located near the center 
of the earth. Such a field is nonuniform, that is, the magnetic lines of force are 
curved and the field strength varies from point to point, decreasing with in- 
creasing distance from the center of the dipole. For a given distance the field 
strength is greatest at the poles and weakest at the geomagnetic equator. The 
axis of the earth's dipole (geomagnetic axis) lies at an angle of 11° from the 
earth's axis of rotation. The midpoint of the dipole does not coincide exaclK 
with the center of the earth. 

Using the earth as a magnet to analyze cosmic rays 


atmosphere. If the primary rays are electrically charged particles, 
they will be deflected by the earth's field in one direction or the 
other depending on whether their charge is positive or negative. 
If they are neutral particles or photons, they will pass through the 
field undeflected. And although the earth's magnetic field is much 
weaker than the fields produced by laboratory electromagnets, its 
enormous extent more than makes up for its weakness. Hence, large 
deflections are to be expected even for charged primary particles 
of very great energy. 

Magnetic rigidity 

To make the discussion in this chapter more precise, it will be 
useful to review very briefly some of the basic facts concerning the 
motion of charged particles in magnetic fields. A moving charged 
particle in a magnetic field experiences a deflecting force that acts 
at right angles to both the magnetic field and the direction of 
motion. If the field is uniform and perpendicular to the direction 
of motion, this continuous sideways thrust 1 causes the particle to 
move in a circle (Fig. 5-2). Now for particles of a given charge, the 
radius R of this circle is inversely proportional to the magnetic 
field strength B and directly proportional to the momentum of the 
particle. 2 

In other words, the product BR (magnetic field times radius) 
is directly proportional to the momentum. For particles of the 
same momentum but with different amounts of charge, on the other 
hand, the product BR is inversely proportional to the charge. A 
multicharged particle experiences a stronger side thrust than a 
singly charged particle, and its trajectory is therefore deflected 
into a circle of smaller radius. 

The product BR is known as the magnetic rigidity of the 
particle. The usual unit for measuring magnetic fields is the gauss; 

1 Generally called the Lorentz force in physics, after the Dutch physicist 
Hendrik Lorentz. 

* See Appendix F. 


Cosmic Rays 



Fig. 5-2 The deflecting force experienced by a positively charged particle 
moving through a uniform magnetic field causes the particle to travel in a 
circle. The direction of the force is given by the following right-hand rule: If 
the fingers of the right hand point in the direction of the Geld B and the thumb 
points in the direction of particle motion v, the force F points out of the palm. 
Since the deflecting force must be directed toward the center of the circle to 
counteract the centrifugal force, a positively charged particle must move 
clockwise with respect to an observer lined up with the magnetic field (a). 
For the same directions of B, the deflecting force experienced by a negatively 
charged particle is directed toward the center when the particle moves counter- 
clockwise (b). For any given set of conditions, reversal of the field, charge, or 
direction of motion reverses the direction of the force. Also, the radius of the 
circle describing the motion of the particle decreases as the momentum of the 
particle decreases and as the strength of the magnetic field increases. 

the earth's magnetic field is a few tenths of a gauss at sea level ; 
strong electromagnets produce fields of the order of 10 4 gauss. If B 
is measured in gauss and R in centimeters, the magnetic rigidity 
BR is measured in gauss cm. The plot in Fig. 5-3 shows the 
magnetic rigidity of electrons and protons as a function of their 
kinetic energy. Note that for sufficiently large values the kinetic 
energy, in electron volts, is simply 300 times the rigidity: 

E (eV) = 300 BR (gauss cm) 
This equation holds, in fact, for all singly charged particles whose 

Using the earth as a magnet to analyze cosmic rays 


kinetic energy is large compared to their rest energy ; the factor 300 
depends on the particular choice of units. 1 For a high-energy par- 
ticle with Z elementary charges, however, the magnetic rigidity 
is Z times smaller than for a singly charged particle of the same 
energy. In this case the relation between energy and magnetic 
rigidity is 

E/Z = 300 BR 

Returning to the question of how the earth's magnetic field 
affects the motion of charged particles, consider a particle that, 

1 See Appendix F. 

10" p 

10' r 

10° r 


CQ 10 5 

10< r 



1 i | .... 

1 1 1 | 1 III 

1 1 1 | ITH 

' ' ' |— 

I I , |.-T 














: ^ 




>V _ 



r / 


i /. 

1 1 ■ 1 1 in 

— i — ' i 1 1 in 

' l l 1 i nr 



10 s 



E, .v 

Fig. 5-3 Magnetic rigidity BR as a function of kinetic energy E for protons 
and electrons. Both E and RR are plotted on a logarithmic scale. For suf- 
ficiently large energies both curves approach the straight line representing 
the equation E = 300 BR. 


Cosmic Rays 

under the action of this field, circles the earth at the geomagnetic 
equator. (Here we disregard energy losses in the atmosphere.) In 
the first place, to keep the particle in this particular orbit, the force 
exerted by the magnetic field must point toward the center of the 
earth. From the rule for determining the direction of the force 
(Fig. 5-2) it is clear that the particle must move from east to west 
if it is positive and from west to east if it is negative (Fig. 5-4). In 
the second place, the magnetic rigidity of the particle must equal 
the product of the radius of the earth (R = 6.38 X 10 8 cm) and the 
field strength at the equator (B = 0.32 gauss): 

BR = 0.32 X 6.38 X 10" = 2 X 10 8 gauss -cm 
As Fig. 5-3 shows, a magnetic rigidity of 2 X 10 8 gauss cm (for 
both electrons and protons) corresponds to a kinetic energy in the 
neighborhood of 60 BeV. Consequently, charged particles with 
energies of this order or less must be strongly deflected by the 
earth's magnetic field. 

The search for a latitude effect 

In 1930 the notions about the possible effects of the earth's mag- 
netic field upon cosmic rays were still rather nebulous from the 

Using the earth as a magnet to analyze cosmic rays 


Fig. 5-4 A positively charged particle circling the earth under the influence 
of the earth's magnetic field moves from east to west. The magnetic field B 
points toward the geomagnetic north pole, and the deflecting force F toward 
the center O of the earth. 

point of view of both theory and experiment. It was thought that if 
primary cosmic rays were electrically charged, they would somehow 
be channeled toward the poles along the magnetic lines of force. 
At high latitudes, then, the cosmic-ray intensity should be greater 
than in the equatorial regions. In 1927 the Dutch physicist J. Clay 
had measured the cosmic-ray intensity, by using an ionization 
chamber, on a trip from Leiden to Java and had found a decrease 
of several per cent in the vicinity of the Suez Canal. On the other 
hand, Millikan and his associates in 1928 had not discovered any 
significant change between Bolivia (19° S latitude) and Pasadena, 
California (34° N latitude). In 1930 Millikan again found prac- 
tically no change between Pasadena and Churchill, Canada (59° N 
latitude). That same year, Bothe and Kohlhorster failed to detect 
any variation with latitude in the North Sea and the Swedish 
physicist Axel Corlin thought he had detected a slight latitude 
effect in the Baltic Sea. 

Needless to say, the existence of a latitude effect seemed very 
doubtful. In any case the effect, if present at all, was small and 
could not be ascribed with any confidence to the earth's magnetic 
field. It was argued, with good reason, that the different conditions 
of the earth's atmosphere at different geographic locations could 
of themselves produce significant changes in the cosmic-ray inten- 
sity at sea level. To Millikan and his collaborators, the absence of a 
large latitude effect was proof that primary cosmic rays were not 
electrically charged. 

Theory of geomagnetic effects 

Much of the theoretical background for the correct treatment of the 
effects I have been discussing was already available at the time. 
The Norwegian geophysicist Carl Stormer, in an effort to explain 
those striking luminous displays of the upper atmosphere known as 
northern lights or aurorae, had made a detailed study of the motion 
of charged particles in the magnetic field of a dipole, to which, as I 


Cosmic Rays 

mentioned above, the magnetic Geld of the earth bears a close 
resemblance. Stormer believed aurorae were caused by charged 
particles coming from the sun at times of increased solar activity. 
From painstaking numerical calculations over a period of years, he 
had derived a large number of possible trajectories for these par- 
ticles. More importantly, he had also established some of the 
general characteristics of the trajectories. 

Although Stbrmer'B basic assumption was correct, his theory 
proved inadequate. One of the reasons is that auroral particles have 
comparatively small energies and are therefore strongly influenced 
by the weak magnetic and electric fields present in interplanetary 
space, fields that were still unknown when Stormer was developing 
his theory. (Present understanding of conditions prevailing in 
interplanetary space indicates that aurorae are much more complex 
phenomena than Stormer had any ground to suppose, phenomena 
so complex in fact, that they have resisted all attempts at detailed 
explanation ; see Chap. 14.) 

Cosmic-ray particles, on the other hand, have far greater 
energies than auroral particles. No valid reason exists today — any 
more than one existed some thirty years ago — for doubting that 
their trajectories closely resemble those computed by Stormer. 
Thus Stormer's theory, developed originally for a different purpose, 
found its most useful application in cosmic-ray studies. This appli- 
cation was not immediate, however, because the step from the 
problem of aurorae to the problem of cosmic rays was not an 
obvious one. 

Stormer, as I said, had computed the trajectories of particles 
with different magnetic rigidities approaching the earth from a 
definite direction, that of the sun. His aim was to discover over 
which areas of the earth, and in which directions, these particles 
entered the atmosphere. But cosmic rays were known to come from 
all directions. A study of their geomagnetic effects by a similar 
method would have required the computation of a far greater 
number of trajectories. This was nearly impossible, especially be- 
fore the invention of electronic computers. 

Using the earth as a magnet to analyze cosmic rays 


I became personally interested in geomagnetic effects in 1930. 
While studying the papers of Stormer, I realized that if one would 
only ask the right questions, one could readily obtain from 
Stormer's work answers that went a long way toward a solution of 
the problems of concern to cosmic-ray physicists. Let me explain. 

To formulate the questions correctly one ought, in the first 
place, to look backward rather than forward in time. Instead of 
following a hypothetical charged particle of given rigidity in its 
motion from outer space toward the earth, it was easier to consider 
a hypothetical particle arriving at a given point P near the earth's 
surface in a given direction (Fig. 5-5) and then to follow its 
trajectory in reverse through the earth's magnetic field. 

Fig. 5-5 Trajectories of two hypothetical particles of different magnetic 
rigidity arriving at the same point P from the same direction (schematic). 
Trajectory a, when traced backward, escapes to inGnity; it is an allowed 
trajectory. Trajectory 6, when traced backward, returns to the earth; it is a 
forbidden trajectory. 


Cosmic Rays 

Two possibilities existed. Either the trajectory, when traced 
backward, would escape from the neighborhood of the earth with- 
out ever crossing the earth's surface or it would come back to the 
earth. In the first instance, the hypothetical particle could have 
come from outer space and thus might have been a cosmic-ray 
particle. In other words, the given direction at P was an allowed 
direction of arrival for cosmic-ray particles of the specified magnetic 
rigidity. In the second case, the hypothetical particle could not have 
come from outer space, and thus was certainly not a cosmic-ray 
particle. In other words, the given direction at P was a forbidden 
direction of arrival for cosmic-ray particles of the specified magnetic 

Cosmic-ray physicists were not primarily interested in com- 
puting actual trajectories. For them the basic problem was to 
determine which directions were allowed and which forbidden. 
And here Stormer's theory provided a major break. Stormer had 
shown that there exists a special class of trajectories (bounded 
trajectories) that remain forever in the vicinity of the dipole. 
Bounded trajectories are certainly not possible trajectories of 
cosmic-ray particles. 

Stormer had also worked out a simple formula by means of 
which, from the position, the direction of motion, and the magnetic 
rigidity of a given particle, it was possible to determine whether or 
not the particle was moving along a bounded trajectory. By using 
this formula, I arrived at the following conclusion: For each point 
on the earth and for positive particles of any given magnetic 
rigidity there exists a cone (Stormer cone), with the axis pointing 
toward the east, such that all directions within the cone corre- 
spond to bounded trajectories and are therefore forbidden direc- 
tions (Fig. 5-6a). If the particles are negative instead of positive, 
the forbidden directions fill a Stormer cone that is the mirror image 
of the first with respect to the meridian plane (that is, the vertical 
plane on the "north-south line"; see Fig. 5-66). 

The graphs in Fig. 5-7 show how the half angle of the Stormer 
cone (angle a in Fig. 5-6a and 6) varies with magnetic rigidity at 

Using Ihe earth as a magnet to analyze cosmic rays 


different geomagnetic latitudes. For example, at X = 20°, a becomes 
180° for BR = 2.8 X 10 7 gauss • cm, which means that all directions 
are forbidden for particles with rigidity equal to or less than 
2.8 X 10' gauss cm. The half angle becomes 0° for BR = 7.8 X 10 7 
gauss cm, which means that, for particles with a rigidity greater 
than this value, the Stormer cone disappears (in other words, there 

Fig 5-6 Stormer cones for positive and negative particles of the same mag- 
netic rigidity. The plane shown in (a) and (6) is the horizontal plane at the 
point of observation 0. For positive particles (a), all directions east of the cone 
(such as AO) are forbidden; directions west of the cone (such as BO) may be 
allowed. For negative particles (6), AO is a forbidden direction; BO may be an 
allowed direction. Angle a is the half angle of aperture of the cones. 


Cosmic Rays 











- 1 — 

— i — 



I — i — 

I — ' — 

























= 0° 


— 1 








■ l 


8 10 12 

BR, 10' gauss* cm 





Fig. 5-7 Semiaperture of Ihe Stormer cone (angle a in Fig. 5-6) as a function 
of magnetic rigidity BR, in units of 10' gauss- cm, for different geomagnetic 
latitudes X. 

are no directions corresponding to bounded trajectories). At the 
same latitude, a = 90° for BR = 3.9 X 10 7 gauss cm, which 
means that for this rigidity the boundary of the Stormer cone is 
coincident with the meridian plane. 

These results were not yet complete. Although they stated 
that all directions within the Stormer cone correspond to bounded 
trajectories and are therefore forbidden, they did not specify which 
directions outside the cone are allowed. For, although directions 
outside the cone correspond to "unbounded" trajectories, some of 
these trajectories, when traced backward, may still cross the earth 
before escaping into space. Nor was it clear how the cosmic-ray 
intensities in the allowed directions varied with respect to one 

Using the earth as a magnet to analyze cosmic rays 


Nevertheless, it was already possible to draw some interesting 
conclusions. By far the most important one was the prediction of 
an asymmetry in the intensity distribution of cosmic rays. Indeed, 
since in the case of positive particles the forbidden Stormer cones 
point toward the east, positively charged primary cosmic rays 
should arrive in smaller numbers from the eastern than from 
the western regions of the sky. But if primary cosmic rays are 
negatively charged, the Stormer cones point to the west, and the 
situation is reversed. This predicted asymmetry became known 
as the east-west effect. 

It may be useful, as an illustration of these general results, to 
consider in detail the special case of trajectories in the plane of the 
geomagnetic equator. These are particularly simple trajectories, 
since they are the only ones that remain forever in one plane. 
Several equatorial trajectories, all corresponding to particles with 
a magnetic rigidity of 7 X 10' gauss cm, are shown in Fig. 5-8a 
and 6 [positive in (a), negative in (6)1. They are drawn through the 
same point on the geomagnetic equator of the earth. In Fig. 5-8a 
all trajectories striking the earth from directions east of the dotted 
line cross the earth's surface over and over again; they are for- 
bidden trajectories, and in fact they are bounded trajectories. On 
the other hand, all trajectories striking the earth from directions 
west of the dotted line, when traced backward, go directly to 
infinity and are therefore allowed directions. The dotted line itself 
is the intersection of the Stormer cone with the equatorial plane. 
Figure 5-86 is the mirror image of Fig. 5-8a. 

The questions left open by this analysis still concerned me. 
I discussed the problem with Enrico Fermi, who pointed out that 
at least one of the gaps could be easily filled. By using a general 
theorem of mechanics known as the Liouville theorem, he proved 
to me that, since cosmic rays were supposed to be distributed 
uniformly in all directions at large distances from the earth, then- 
intensity near the earth should be the same in all allowed directions. 
Thus the only remaining problem was to find out which of the 
directions outside the Stormer cone were allowed and which 


Cosmic Hays 

This problem was attacked with great vigor in the following 
years, particularly by Georges E. Lemaitre of Belgium, Manuel S. 
Vallarta of Mexico, and their students. In their work these 
scientists used a mechanical computer, known as the differential 
analyzer, that was developed by Vannevar Bush at the Massa- 
chusetts Institute of Technology. They did find that some of the 
directions outside the Stormer cone were forbidden. In most cases, 
the corresponding trajectories were not bounded but, when traced 
backward, crossed the earth before going to infinity. 

To relate in detail the results of these studies would take the 
discussion too far afield, but one of the most significant findings 

Fig. 5-8 Trajectories of positively charged particles (a) and of negatively 
charged particles (6) in the plane of the geomagnetic equator. [The magnetic 
field points out of the drawing. According to the right-hand rule (Fig. 5-2), the 
trajectories must curve in the directions shown.] In both cases the magnetic 
rigidity is 7 X 10' gauss -cm. For this magnetic rigidity there are positive 
particles coming from infinity and striking the earth from all directions between 
32° E and 90° W of the vertical (examples 1 and 2 in illustration). However, in 
order to strike the earth from a direction at more than 32° E of the vertical, 
a particle would have to come from some point on the earth's surface (example 3 
in (o)]. Thus, 1 and 2 are examples of allowed trajectories, 3 is an example of 
forbidden trajectory, and it follows that positive particles arrive more abun- 
dantly from the western than from the eastern regions of the sky. Another way 
of clarifying the east-west effect is to consider particles, such as 2, that 
approach the earth from the east but are then deflected by the magnetic field 
so that when they actually strike the earth, they appear to come from the west. 
(The same considerations, with the directions east and west interchanged, 
apply to negative particles.) 

Using the earlli as a magnet to analyze cosmic rays 


appears in Fig. 5-9. Each curve in the figure describes the effect of 
the magnetic field on the total flux at the top of the atmosphere of 
primary particles of specified rigidity. These curves afford quan- 
titative predictions about the latitude effect of cosmic-ray particles 
with different magnetic rigidities. They show, for example, that 
particles with a magnetic rigidity of 3.2 X 10 7 gauss-cm are 





1.28 X 10 B 

9.8 A y^'. 

/ *~* 



/ * 
1 * 

1 ° 

1 * 
1 °° 




II — 

0° 10° 20' 30' 40° 

Geomognetic latitude 




Fig. 5-9 Latitude effect on cosmic-ray intensity, according to the theoretical 
studies of Lemaitre and Vallarta. Each curve refers to particles of a particular 
magnetic rigidity, and gives the total number of such particles reaching the 
top of the atmosphere as a function of geomagnetic latitude. The particles arc 
assumed to approach the earth in equal numbers from all directions. The num- 
ber of particles of each magnetic rigidity that would strike the atmosphere if 
there v.ere no magnetic field is arbitrarily taken to be 100. Cosmic-ray particles 
with a magnetic rigidity of 3.2 X 10' gauss cm are completely excluded from 
an equatorial belt extending from about 12° N to 12° S geomagnetic latitude. 
The maximum number (100) arrive at latitudes greater than about 43° N 
and S. At a latitude of 32° N and S the magnetic field reduces the intensity by 
approximately one-half. (From a paper in The Physical Review, vol. 43, p. 87, 


Cosmic Rays 

completely excluded from an equatorial belt extending from about 
12° N to 12° S geomagnetic latitude; that the flux of these par- 
ticles increases gradually between 12° and 43°, north and south; 
and that above 43° it reaches a constant value equal to the flux 
that would be observed in the absence of any magnetic field. 

Discovery of the east-west effect and further 
measurements of the latitude effect 

The theoretical work on geomagnetic effects was accompanied, and 
indeed stimulated, by extensive experimental work in which scien- 
tists from many different countries participated. The most ambi- 
tious program was the one undertaken in 1930 by A. H. Compton, 
who, with the help of many collaborators, carried out a world-wide 
survey of cosmic-ray intensities both at sea level and at high 
altitudes on mountains. 

For my own part, I concentrated my efforts on the east-west 
effect. There were two main reasons for this preference. One was 
that the east-west effect, to my mind, provided a surer test of 
magnetic influence than the latitude effect, because it did not 
involve a comparison of cosmic-ray intensities at different loca- 
tions. Variations in the conditions of the atmosphere from one 
region of the earth to another, which could conceivably simulate a 
magnetic latitude effect, could not produce an east-west asym- 
metry. The second reason, of course, was that if cosmic rays were 
electrically charged, the east-west effect would also reveal the sign 
of the charge. 

In 1931, at the Arcetri laboratory, I made my first attempt to 
detect an east-west effect. The type of instrument I used for this 
purpose later became known as the cosmic-ray telescope. This instru- 
ment consists simply of two or more G-M counters in coincidence 
arranged with their centers on a straight line called the axis 
(Fig. 5-10). Only particles coming from directions near the axis 
can traverse all counters and thereby produce simultaneous dis- 
charges, or coincidences. By changing the orientation of the tele- 

Using the earth as a magnet to analyze cosmic rays 


scope, it then becomes possible to compare the numbers of cosmic- 
ray particles arriving from different directions. The field of view of 
the telescope depends, of course, on the size and separation of the 
counters. Figure 5- 10a shows a telescope with a wide field of view; 
Fig. 5-106, a telescope with a narrow field of view. In the experi- 
ment, I pointed the G-M telescope alternately to the east and to 
the west of the geomagnetic meridian (Fig. 5-11), but I was unable 
to detect any significant difference between the counting rates 
recorded in the two directions. Then in 1933, as the first concrete 
evidence for the existence of a latitude effect was beginning to 
appear, with some indication that the effect might be caused by 
the earth's magnetic field rather than by meteorological factors, the 
east-west effect was discovered in three separate experiments. 
Thomas H. Johnson, working in Mexico City (29° N geomagnetic 
latitude, 2,250 meters above sea level), and Luis W. Alvarez in 
collaboration with Compton, working at the same location, found 
the intensity of cosmic rays to be greater from the west than from 
the east. The difference was approximately 10 per cent with the 


Fig. 5-10 Geiger-MUUer telescopes consisting of two G-M counters in 
coincidence, (a) A telescope with a wide field of view; (6) a telescope with a 
narrow field of view. The dotted line through the centers is the axis of the 


Cosmic Rays 


Fig. 5-11 Experimental study of the cast-west effect: a cosmic-ray telescope 
pointing (a) to the east and (ft) to the west of the geomagnetic meridian. In 
both cases, the telescope is inclined at the same angle 2 from the vertical. 

cosmic-ray telescope set at an angle of 45° to the vertical. A few 
months later Sergio De Benedetti and I, working in Asmara, 
Eritrea (1 1° N geomagnetic latitude, 2,370 meters above sea level), 
found an excess of 26 per cent in the west direction, again with the 
telescope set at 45° to the vertical. 

It was thus clear that a portion, possibly all, of the primary 
cosmic radiation consisted of positively charged particles. This was 
a most unexpected result, for the common belief among proponents 

Using the earth as a magnet to analyze cosmic rays 


of the corpuscular hypothesis had been that the primary cosmic 
radiation consisted of electrons. Had the east-west effect been 
discovered one year earlier, we would have concluded that cosmic 
rays must be protons or heavier nuclei, since these were the only 
positively charged particles known at that time. And it so happens 
we would have been right. But in 1932 a new positive particle had 
been discovered — the positive electron or positron (Chap. 6). Thus 
we had to leave the question of the nature of primary cosmic rays 
in abeyance until new evidence developed. 

In the following years a wealth of experimental data on the 
east-west effect, and particularly on the latitude effect, became 
available. In 1936 Compton summarized the existing information 
in the form of a graph, which is reproduced in Fig. 5-12. The curves 
in the graph are lines of equal cosmic-ray intensity, or isocosms, as 
Compton called them. The same graph also shows the geomagnetic 
equator and the geomagnetic parallels at 50° N and 50° S. As the 
figure makes evident, the isocosms follow geomagnetic parallels 
more closely than they do geographic parallels. This fact supported 
the view that the intensity variations of cosmic radiation over the 
earth's surface were due primarily to the terrestrial magnetic field 
rather than to atmospheric factors. 

The possibility of some atmospheric influence, however, could 
not be ruled out. Indeed, in 1937 Compton and his collaborators 
obtained positive evidence of a slight decrease in the intensity of 
cosmic rays at sea level as the temperature of the atmosphere 
increased. Since the average temperature of the atmosphere is 
greater near the equator than near the poles, this temperature 
variation tended to produce an apparent latitude effect. Precise 
measurements showed that about one-third of the latitude effect 
found at sea level was actually due to the temperature variation; 
the earth's magnetic field accounted for about two-thirds of the 
total effect. 

But the intensity of cosmic radiation varies not only with 
latitude but also with longitude, as Compton 's graph shows (Fig. 
5-12). For example, the intensity along the geomagnetic equator 





O u 

f I 


S — 
5 - 

O, r 



& § 










* 1 

1 1 

3 J3 


■8 £ 

i 1 

£ en 

01 « 

o Z 
"■3 o 




1 s 

£ & 
S -a 

8 ?? 



■a -a 

5 1 S 



■a e 


Using the earth as a magnet to analyze cosmic rays 


falls to a minimum in the Indian Ocean. This observation, too, fits 
the picture. As I mentioned at the beginning of tins chapter, the 
center of the earth's magnetic dipole does not coincide exactly with 
the geographic center. It is displaced toward the Indian Ocean, so 
that the magnetic field there is stronger and produces a more 
pronounced decrease in the cosmic-ray intensity. 

The effect of the atmosphere 

Thus far, I have ignored the atmosphere, and the time has come to 
inquire how its presence might affect our discussion. As I pointed 
out in the preceding chapter, at the time when the investigations 
related here were made, there were already good reasons for 
believing that the local radiation observed in the atmosphere 
consisted largely of secondary particles. What, then, was our 
justification for interpreting the observatior.3 on the basis of a 
theory that actually applied to the primary radiation arriving at 
the top of the atmosphere from outer space? 

With regard to the latitude effect the answer was clear. Since 
the number of secondary particles observed under the atmospheric 
blanket would increase or decrease as the number of primary rays 
falling upon the atmosphere increased or decreased, secondary rays 
would exhibit a latitude effect similar to that of the primary rays. 
The effect would not, however, be quite as pronounced, because the 
primary particles that show the largest latitude variations are those 
of lowest energy, and these are likely to produce fewer secondary 
particles in the atmosphere. 

In this connection, it is interesting to consider the measure- 
ments at high altitudes on mountains and by means of balloon 
that were made in the early 1930s by the teams working with 
Compton and with Millikan. These measurements showed that the 
latitude effect became more pronounced as the altitude increased 
(Fig. 5-13). This result was easy to account for. Under a smaller 
atmospheric thickness (that is, at high altitudes) the electroscopes 


Cosmic Rays 

or the G-M counters could detect (either directly or through their 
secondary rays) primary particles of lower energy, which were 
more strongly influenced by the earth's magnetic Geld. 

The interpretation of the east-west effect, on the other hand, 
involved different considerations. Tf the particles produced in the 
collisions of primary rays with atoms and molecules went off at 
wide angles to the primary direction, the secondary radiation in 
the atmosphere would be distributed more or less uniformly in all 
directions, irrespective of the direction from which the primary 
radiation came. In that case, we could hope to observe an east-west 
effect only by sending our instruments clear of the atmosphere, 
where the primary cosmic radiation arrived undisturbed. Nonethe- 
less, we had detected an east-west effect under a considerable 
atmospheric thickness, where presumably most of the observed 
rays were secondary. Thus we concluded that the secondary rays 
produced in the atmosphere must be strongly "lined up" in the 
direction of the primary radiation. 

This conclusion received support from another experimental 
observation. Measurements with the cosmic-ray telescope revealed 
a sharp maximum of intensity in the local radiation when the tele- 
scope was pointed straight up. The maximum could be explained 
as an effect of atmospheric absorption, because the radiation trav- 
ersing the atmosphere vertically passed through a smaller amount 
of matter than the radiation traversing the atmosphere in any 
other direction. Such an explanation, however, would hold true 
only if the secondary rays traveled in more or less the same direc- 
tion as the primary rays that produced them. 

Using the earth as a magnet to analyze cosmic rays 


20 40 60 

Geomagnetic latitude 

Fig. 5-13 Cosmic-ray intensity as a function of geomagnetic latitude at three 
different altitudes. The intensity is measured by the number of ion pairs 
produced by cosmic rays in 1 cm* of air at normal temperature and pressure. 
(From A. H. Compton. The Physical Revietv, vol. 43, p. 387, 1933.) 

Positrons and the 


of energy 



Through the yeare cosmic-ray physicists have been faced with two 
distinct but closely related problems. The first, of course, concerns 
the nature of cosmic rays. The second is: What kinds of particles 
exist in nature, and how do particles of very high energy behave? 
It would have been comparatively easy for the physicists to solve 
the first problem if they had known the answer to the second, and 
vice versa. What made life difficult was the unavoidable necessity 
of having to deal with both problems simultaneously. The story to 
be told in this and the next several chapters is a fine example of an 
enterprise in which skillful experimentation, bold and critical 
reasoning in the interpretation of experimental results, and imagi- 
native theoretical thinking were brought to bear upon a most 
intricate scientific task. The successful conclusion of these efforts 
brought not only an answer to the cosmic-ray puzzle but also a 
dramatic advance over a wide front of physics. 



Cosmic Rays 

Positrons and tlte materialization of energy 


The story begins with the local cosmic radiation. In the experi- 
ments of Bothe and Kohlhorster the coincidence rate had dropped 
24 per cent when 4.1 cm of gold was placed between the counters. 
On the other hand, my own experiments had shown that, after 
passing through some thicknesses of lead, cosmic-ray particles 
became much "harder" and that a large proportion of them were 
capable of traversing lead thicknesses of the order of meters. Other 
measurements confirmed these results (Fig. 6-1). Therefore, the 
local cosmic radiation contained two distinct groups of particles: a 
"soft" group, responsible for the initial fast drop in the absorption 
curve of Fig. 6-1, and a "penetrating" group, responsible for the 
flat "tail" of the curve. 

Moreover, as we had learned, at least some of the cosmic-ray 
particles produced the complex secondary processes called showers. 
However, we had not yet found out what the various kinds of 
secondary particles were. As for the showers, they were thought to 
result from the disintegration of atomic nuclei, but this assumption 
remained to be tested. 


«- 75 

£ £ so 

■o ? 


o .9, 

U 25 


5 S 

Lead, cm 

5 1( 



Fig. 6-1 Soft and penetrating groups of cosmic-ray particles (at sea level) 
account respectively for the sharp initial drop and the natter tail of the absorp- 
tion curve (see text). The horizontal scale gives the total thickness of the lead 
absorber between counters. The vertical scale gives the coincidence rata, in 
arbitrary units. The curve represents data obtained from experiments per- 
formed by the author in 1932. 

The discovery of the positron 

Even before the discovery of showers, Carl D. Anderson, in 
Millikan's laboratory at the California Institute of Technology, 
had started an experimental program that, together with the one 
begun a little later by Blackett and Occhialini, was to provide a 
partial answer to our questions. The major result of this work was 
the discovery of the positive electron, or positron, and of the strange 
circumstances attending its birth and its disappearance. 

In his experiments Anderson used a cloud chamber placed in 
the field of a powerful electromagnet. At a maximum field strength 
of 24,000 gauss he was able to measure the magnetic deflection of 
tracks whose radius of curvature was as great as 7 meters and 
whose magnetic rigidity was therefore roughly 1.7 X 10' gauss -cm 
(that is, 24,000 gauss X 700 cm). The corresponding kinetic ener- 
gies — about 5 BeV for particles with the mass of an electron and 
about 4 BeV for particles with the mass of a proton (Fig. 5-3) — 
were several hundred times greater than any energy that previous 
instruments had been capable of measuring. It soon became appar- 
ent that the particles of the local cosmic radiation had a wide 
range of energies extending well beyond 1 BeV. In addition, about 
half their trajectories bent to the right and half to the left. As- 
suming the direction of travel to be downward in every case, 
Anderson concluded that positively and negatively charged par- 
ticles were about equally abundant in the local cosmic radiation. 

At first Anderson thought the negative particles were electrons 
and the positive particles protons. As the experimental data 
became more precise, however, he was obliged to abandon this 
interpretation; for the cloud chamber also made it possible to 
estimate the ionizing power of the particles from the number of 
droplets along their tracks. Now, the ionizing power decreases 
with increasing velocity (Chap. 2). For a given magnetic 
rigidity, the velocity of a lighter particle is greater than that 
of a heavier particle. Therefore, the lighter particle has a smaller 
ionizing effect. 


Cosmic Rays 

To illustrate this point, the graph in Fig. 6-2 gives the ion 
density as a function of rigidity for electrons and protons. The 
curves (similar to the ones in Fig. 2-2, which give ion density as a 
function of kinetic energy) show that electrons and protons with 
sufficiently small magnetic rigidities produce very different densities 
of ionization. Consequently, if Anderson's positive particles were 
protons, their tracks should have been denser than the tracks of the 
negative particles. 

Contrary to expectation, the ionization densities were prac- 
tically the same, even for magnetic rigidities in the range below 10 8 
gauss • cm, where protons would have ionized at least six times more 
heavily than electrons. It was tempting to accept these results as 
evidence for the existence of a hitherto unknown particle with a 
positive elementary charge and a mass close to that of the electron. 
But that was such a daring conclusion that all possible sources of 
error had to be critically examined. Hence, Anderson gave serious 
consideration to the possibility that the particles he thought were 
positive particles moving downward might, in reality, be negative 
electrons moving upward. To test this alternative explanation, 
Anderson partitioned his cloud chamber into an upper and lower 
section by means of a lead plate. A particle traversing the plate 
would suffer an energy loss and thus emerge with a correspondingly 
reduced rigidity, thereby disclosing its direction of motion. 

The results obtained with the new arrangement revealed par- 
ticles moving upward as well as downward, fully justifying Ander- 
son's misgivings. However, as the cloud chamber photographs also 
proved beyond any doubt, some of the positive particles had nearly 
the same mass as the electron. 1 Curiously enough, the most con- 
vincing evidence came from the picture of a positive particle 
moving upward; in the absence of the lead plate, it would un- 
doubtedly have been mistaken for an ordinary negative electron 

1 Accurate measurements of the radii of curvature and the ion densities 
of a number of tracks enabled Anderson to make a fairly precise estimate of 
the mass. According to his results, this mass did not differ by more than 
20 per cent from that of the ordinary negative electron. 

: i 













— a. — 






/ ' 








i - 


1 . 






Cosmic Rays 

moving downward! (See Fig. 6-3.) The direction of motion of the 
particle is obvious from the fact that the portion of the track above 
the plate has a considerably smaller radius of curvature than the 
portion below the plate. The magnetic deflection indicates a 
positively charged particle. The magnetic rigidity changes from 
2.1 X 10 6 gauss cm to 7.6 X 10* gauss cm as the particle trav- 
erses the plate, and the ionization density appears to be about 
minimum on either side. 

All these results are easily understood if the particle has a mass 
close to that of an electron. But a proton with a magnetic rigidity 

Fig. 6-3 The positron, or positive electron, was identified as the particle that 
entered the cloud chamber from below and produced the track curving sharply 
to the left after traversing the lead plate. The photograph, taken by Anderson 
in 1932, definitely established the existence of positrons. (From a paper in 
The Physical Review, vol. 43, p. 491, 1933.) 

Positrons and the materialization of energy 


of 7.6 X 10* gauss cm would produce a density of ionization much 
greater than the density observed. Moreover, it would have a 
range of only 5 mm in the gas of the chamber, whereas the observed 
length of the track is more than 50 cm. 

Dirac's theory and the origin of positrons 

At the time of Anderson's work in the United States, Blackett and 
Occhialini, in England, were carrying out their experiments with 
the counter-controlled cloud chamber. These experiments, besides 
revealing the great complexity of cosmic-ray interactions, produced 
other results of crucial importance. Tn the first place, Blackett and 
Occhialini were able to confirm Anderson's discovery of the posi- 
tron, announced a few months earlier. In the second place, they 
found that all shower particles deflected by the comparatively 
weak magnetic fields used in their experiments (usually from 2,000 
to 3,000 gauss) were either negative or positive electrons. And, in 
some showers at least, the numbers of positive and negative par- 
ticles were comparable. 

Blackett and Occhialini had also taken many pictures with the 
cloud chamber partitioned by a lead plate. The pictures often 
showed two or more showers originating simultaneously from the 
plate and the chamber walls. The frequent occurrence of multiple 
showers was quite surprising. It looked as if particles generated in 
showers were more likely to produce secondary showers than were 
the average cosmic-ray particles found in the atmosphere. Last of 
all, in a number of pictures Blackett and Occhialini found showers 
that appeared to originate in the plate without any ionizing track 
entering the plate from above. 

Such showers, therefore, were not produced by electrons or by 
any other charged particles. It became necessary to postulate the 
existence of neutral rays capable of producing showers. In Blackett 
and Occhialini's view, the neutral rays were probably high-energy 
photons. The experimenters went on to point out that apparently 


Cosmic Rays 

these photons were often to be found among the secondary products 
of showers; for it was not unusual to detect multiple showers with a 
"nonionizing link" (Fig. 6-4). 

One of the members of the Cavendish Laboratory, where 
Blackett and Occhialini did their experiments, was P.A.M. Dirac, 
a young theorist who a few years earlier had developed a new 
theory of the electron. In his theory Dirac sought to combine the 
basic principles of quantum mechanics with the postulates of 
Einstein's relativity theory. To most people Dirac's theory did not 
make much sense: it predicted such obviously absurd things as 
particles with negative mass and negative energy. 

Dirac showed, however, by a rather fanciful argument, that 
his theory could be salvaged if one assumed the existence of positive 
as well as negative electrons. Although he tried to identify the 
positive electron with the proton, his theory stubbornly refused to 
assign to the hypothetical particle a mass different from that of the 

Fig. 6-4 A shower originating in the material above a cloud chamber contains 
a nonionizing link or photon (dashed line), which produces a secondary shower 
in the horizontal metal plate across the chamber. (Traced from a photograph 
published by Blackett and Occhialini, Proceedings of the Royal Society (London), 
vol. A139, p. 699, 1933.] 

Positrons and the materialization of energy 


negative electron. The two particles had to be exactly identical, 
except for the different sign of their charges. According to an 
expression that was to become fashionable later, each was the 
"antiparticle" of the other. Blackett and Occhialini saw in the light 
positive particles of their experiments nothing other than the 
positive electrons required by Dirac's theory. Immediately the 
pieces of the puzzle began to fall into place. 

Why were positrons so rare that they had previously eluded 
observation? The answer, according to Dirac's theory, was that a 
positive electron in matter had a very short life, because, when it 
met a negative electron, the two particles annihilated each other. 
Their mass was changed into energy in agreement with Einstein's 
equation E = mc 2 , and this energy was radiated in the form of 

Why did showers appear to contain equal numbers of positive 
and negative electrons? The answer, again according to Dirac's 
theory, was that a positive electron was always produced together 
with a negative electron. The mass of the two electrons resulted 
from a materialization process (known as pair production) in which 
part or all of the energy of the primary particle was transformed 
into mass. 

What were the nonionizing links found by Blackett and 
Occhialini in their cloud-chamber photographs? They were in all 
likelihood photons, because Dirac's theory predicted that high- 
energy photons were capable of producing pairs of positive and 
negative electrons through the materialization of energy. 

Finally, some physicists had noticed that high-energy y rays 
from radioactive sources were absorbed in matter more rapidly 
than the theory of Klein and Nishina predicted (Chap. 2). The 
"excess" absorption could now be explained by pair production. 
At sufficiently high energy, photons were absorbed not only in 
Compton collisions (to which the theory of Klein and Nishina 
applies), but also in the materialization of photon energy into 
positive and negative electrons. James Chadwick, Blackett, and 
Occhialini in England, as well as Irdne and Frecleric Joliot-Curie in 


Cosmic Rays 

France, tested the pair-production hypothesis experimentally. 
They found that positrons were also produced by 7 rays from 
terrestrial sources, specifically from beryllium bombarded with a 

Electrons, photons, 
and showers 


The discovery of the positron, the experimental test of Dirac's 
theory, and the direct observation of events in which energy is 
transformed into matter and matter into energy are among the 
most striking accomplishments of physics in the twentieth century. 
Nevertheless, these achievements left unanswered many questions 
about cosmic-ray showers. Blackett and Occhialini had found, 
certainly, that showers were often the result of two or more suc- 
cessive collisions. But was this always the case, and how many 
particles (or, rather, how many pairs of positive and negative 
electrons) were produced in each collision? What was the nature of 
these interactions? Did they involve atomic nuclei? If so, were the 
nuclei disrupted? And which particles in the local cosmic radiation 
were capable of initiating showers? The need for new experimental 
data and new theoretical work was greater than ever. 

Experimental properties of showers 

Although the cloud chamber was unsurpassed as an instrument for 
discovering the nature of complex interactions and for analyzing 



Cosmic Rays 

individual events, some of the most significant data on showers and 
the radiation that produced them came from the simpler, cruder 
apparatus of G-M counters out of line (Fig. 4-3). The years follow- 
ing the discovery of showers saw numerous coincidence experi- 
ments performed with G-M counter arrays in many different 
laboratories. I shall not even attempt to give a comprehensive 
account of the new information thus gathered; instead, I shall 
merely mention two early results that were particularly significant 
for the understanding of showers. 

The first result was the so-called shower curve, which describes 
how the coincidence rate varies as the thickness of the absorber 
above the counters increases. The typical shower curve shown in 
Fig. 7-1 was derived from data obtained with the experimental 
arrangement shown in the inset. The material above the counters 
was lead. As the curve shows, when the lead thickness was in- 
creased, the rate of coincidences due to showers coming out of the 
lead increased rapidly at first, reached a maximum between 1 and 2 
cm, and then — quite unexpectedly — decreased rapidly again. 

At the time, many of us had thought the penetrating particles 
found in the atmosphere were the primary cosmic rays themselves. 
Supposedly these particles initiated showers by interacting with 
atomic nuclei. The shape of the shower curve told us we were on 
the wrong track. The penetrating cosmic-ray particles had a known 
mean range of the order of meters in lead; their numbers and their 
energies would not change appreciably in the penetration of a few 
centimeters of the metal. And yet the number of showers coming 
from a layer of lead decreased by roughly one-half when the lead- 
layer thickness was increased from 2 to 5 cm. Clearly, the radiation 
responsible for showers was much more easily absorbed by lead 
than the penetrating corpuscular radiation was. 

The second result had to do with the rate at which showers 
occurred in different substances. In placing above the counters 
layers of lead, iron, and aluminum all having the same mass per 
unit area (several grams per square centimeter), I found an 
approximate shower ratio of 4:2:1 for the three metals. Thus 

Electrons, photons, and showers 


showers were produced much more abundantly in a given mass of a 
heavy element (for example, lead) than in the same mass of a light 
element (for example, aluminum). Other experiments showed that 
heavier elements were also more effective absorbers of the radiation 
responsible for the showers. This peculiar difference in behavior 
between heavy and light elements was a novel feature in high- 
energy physics. The ionization losses of charged particles, as well 
as the occurrence of Compton collisions, were approximately the 
same in different materials when thicknesses were measured in 
grams per square centimeter rather than in centimeters. 



_ 20 



1 1 




1. >l 

— — 

5 cm 

12 3 4 5 6 

Thickness of lead, cm 

Fig. 7-1 Shower curve. The number of coincidences per hour is plotted as a 
function of the thickness of lead above the counters. The experimental arrange- 
ment is shown schematically in the inset. The circles are experimental points. 
i J 'h is figure is based on one appearing in a paper by the author in Zeilschrijt 
fur Physik, vol. 82, p. 151, 1933.) 


Cosmic Rays 

Electrons, photons, and showers 


These experimental observations, in addition to others I have 
omitted, suggested a number of interesting conclusions concerning 
the "shower-producing radiation" and its place in the general 
framework of cosmic-ray phenomena. A concrete and precise 
picture began very gradually to emerge. Fortunately, an important 
theoretical development made it unnecessary to construct the com- 
plete picture piece by piece from purely empirical data. This 
development came about not because physicists had discovered 
any essentially new principle, but rather because, from existing 
theories, they had succeeded in making accurate predictions of the 
behavior of high-energy electrons and photons. 

The theory of Bethe and Heitler 

It was known that fast electrons, when rapidly brought to rest in 
matter, produced photons. (This, in fact, is the manner in which 
X-rays are generated.) It was generally believed, however, that the 
energy lost by electrons through photon emission constituted a 
small fraction of the energy lost through ionization. But in 1934, 
Hans A. Bethe and W. Heitler, then in England, reported the 
results of calculations pointing to a very different conclusion. They 
had considered in detail what happend when a charged particle 
passed near an atomic nucleus and its trajectory was bent by the 
strong electric field associated with the positive electric charge of 
the nucleus. According to the classical theory of electric and 
magnetic fields, any accelerated charge emits electromagnetic 
waves. Since the motion along a curved trajectory is accelerated 
motion, charged particles passing near a nucleus must radiate; that 
is, in the language of quantum physics, the particles must emit 

To compute the effect accurately, Bethe and Heitler had to 
use quantum theory (because they were dealing with individual 
subatomic particles) and relativistic equations (because they were 
dealing with particles moving at almost the speed of light). Their 
main results were as follows: 

1. Radiation losses are enormously greater for light particles 
(such as electrons) Uianfor heavy particles (such as protons). This is 
easily understandable because particles radiate as a result of 
accelerated motion, and for a given force the acceleration is 
inversely proportional to the mass. The greater the mass, the 
smaller the acceleration. 

2. For a given mass per unit area, radiation losses are much 
greater in elements of high atomic number than in elements of low 
atomic number. Rather than go into details, I shall only mention 
that this result is related to the fact that the deflecting force 
experienced by a particle passing near a nucleus is proportional to 
the electric charge on the nucleus. The atomic number Z is the 
number of positive charges (protons) in a nucleus. 1 Thus nuclei 
with higher atomic numbers, that is, with greater electric charges, 
cause a greater acceleration and a correspondingly larger radiation 

3. Radiation losses increase rapidly with energy. Radiation 
losses differ in this respect from ionization losses, which first 
decrease with increasing energy and then become more or less 
constant (Fig. 2-2). As the energy increases, therefore, radiation 
losses will eventually overtake ionization losses. For electrons this 
happens at about 10 MeV in lead and at about 100 MeV in air 
(Fig. 7-2). (The relation of radiation loss to energy cannot be 
explained by any simple argument based on classical physics, 
because it involves relativistic considerations.) 

Bethe and Heitler also reported the rather startling results of 
their calculations on the process of pair production. To repeat one 
of the points in the preceding chapter, Dirac's theory predicted the 
production of pairs of positive and negative electrons by photons. 
Oppenheimer and one of his students had looked somewhat more 
closely into the nature of the actual process. Apparently the most 
likely event was one in which a photon passing through the strong 
electric field of an atomic nucleus suddenly disappeared, giving 

• See Appendix G. 




K) 0.15 




V lonizat 




10* I0 9 

E, eV 

10 10 

Fig. 7-2 Energy losses of electrons through ionization or radiation processes 
(in lead and in air) as a function of their kinetic energy. The vertical scale 
represents the energy loss &E in 1 g/cm* divided by the initial kinetic energy 
E of the electron. The graphs show, for example, that a 100-MeV electron 
(E = 10' eV) traversing 1 g/cm* of lead loses about 1.7 per cent of its energy by 
ionization (fiB/E = 0.017) and, on the average, 16 per cent by radiation 
(&E/E = 0.16). 

birth to a single pair of positive and negative electrons. Bethe and 
Heitler made an accurate computation of this process and found 
the following: 

1. For a given photon energy the probability of pair production 
in layers of different elements with the same mass per unit area 
increases rapidly with the atomic number Z of the element. In this 
case, as in the case of radiation losses of electrons, the dependence 
on atomic number is an effect of the electric charge of nuclei. 
Photons, although they carry no charge, are a form of electro- 
magnetic radiation and will therefore interact with an electric 
field: the stronger the field the more probable the interaction. 
Thus a photon of given energy is more likely to give rise to an 
electron pair when passing through the strong field of a high-Z 

Electrons, photons, and showers 93 

nucleus than it is when passing through the comparatively weak 
field of a low-Z nucleus. 

2. Starting from an energy of 1 MeV, at which pair production 
first becomes possible, the probability of pair production in a given 
thickness of matter first increases rapidly with increasing photon 
energy and then levels off at a nearly constant value. Since the proba- 
bility of Compton collisions decreases steadily with increasing 
energy, it follows that at low energies photons are absorbed mainly 
through the Compton effect and at high energies mainly through 
pair production. The energy at which pair production overtakes the 
Compton effect is about 5 MeV in lead and about 20 MeV in air 
(Fig. 7-3). 

At the time Bethe and Heitler developed their theory it was 
the common belief that all charged particles in the local cosmic 
radiation were either positive or negative electrons. But then the 
fact that many of these particles penetrated as much as 1 meter of 






jj 0.08 








10° 10 7 10" 10" 
E, eV 


Fig. 7-3 The probability that a photon will undergo a Compton collision or 
a materialization event while traversing 1 g/cm 1 of lead or air is plotted as a 
function of photon energy E. The graphs show, for example, that a 10-MeV 
photon (E = 10' eV) has a 1.3 per cent probability of undergoing a Compton 
collision and a 4 per cent probability of disappearing by pair production in 
1 g/cm' of lead. 


Cosmic Rays 

lead appeared to contradict the theory. For if the theory were 
correct, only electrons of absurdly high energies should have been 
capable of traversing such a thick absorber. Moreover, Anderson 
had observed in his cloud chamber charged particles of about 300 
MeV that did not lose nearly as much energy in matter as theory 
predicted. Indeed, Bethe and Heitler, on the basis of the available 
experimental evidence, concluded that "the quantum theory is 
definitely wrong for electrons of such high energy." 

The shower theory 

A short time later, however, there was evidence from a different 
quarter that the theory of Bethe and Heitler, contrary to the belief 
of the authors themselves, was unquestionably correct for electrons 
and photons of very high energy. In fact, the theory provided the 
key to the explanation of cosmic-ray showers. 

The idea on which the explanation rests is really quite simple 
(Fig. 7-4). For example, suppose a high-energy photon — with an 
energy of several BeV, say — enters a block of lead. After traveling 
a short distance (approximately 7 mm according to the theory) it 

Fig. 7-4 Development of a shower in matter through successive events of 
pair production and radiation. Doited lines represent photons, solid lines 

Electrons, photons, and showers 


disappears, giving rise to two electrons (one positive, one negative), 
which between them share the energy of the incident photon. The 
two electrons do not travel far before each of them radiates a 
photon, thereby losing a large fraction of its energy. 

The newly created photons soon materialize into electron 
pairs and the process continues. With each new interaction, two 
particles come from one. Two electrons arise from a single photon; 
one electron and one photon, from a single electron. Correspond- 
ingly, the individual particle energy is, on the average, cut in 
half. As a result, the particles increase in number at first while 
their energy decreasea. 

Eventually, as the original energy is shared among more and 
more of the newly created particles, most of the electrons have so 
little energy that they no longer radiate efficiently and are quickly 
brought to rest by ionization losses. Similarly, more and more of the 
newly radiated photons lack sufficient energy to produce electron 
pairs and are soon absorbed in Compton collisions. The shower, 
grown old, gradually dies out. The photograph in Fig. 7-5 shows an 
actual shower as it develops through a number of brass plates in a 
large cloud chamber. 

This interpretation of cosmic-ray showers was developed al- 
most simultaneously by Homi J. Bhabha and Heitler in England 
and by J. F. Carlson and Oppenheimer in the United States. The 
detailed theory of the cascade process, as the gradual building up of 
showers through pair production and radiative collisions came to be 
known, presented a difficult mathematical problem. Among the 
many scientists who contributed to its solution I should like to 
mention here Lev D. Landau, Igor E. Tamm, and S. Z. Belenky 
of Russia; Hartland S. Snyder, Robert Serber, and Wendell H. 
Furry of the United States; and Bhabha and S. K. Chakrabarty 
of India. 

The problem was of a statistical nature. The exact point at 
which a given photon materializes or a given electron radiates is a 
matter of chance. How the energy of the photon or the electron is 
shared between the two particles produced in a single event is also 


Cosmic Rays 


Fig. 7-5 A shower developing through a number of brass plates 1.25 cm thick 
placed across a cloud chamber. The shower was initiated in the top plate by 
an incident high-energy electron or photon. The photograph was taken by the 
MIT cosmic-ray group. 

Electrons, photons, and showers 


largely a matter of chance. Consequently, showers initiated by 

photons (or by electrons) of a specified energy do not all look alike. 

One may, however, inquire into the average behavior of 

showers. To take an example, Fig. 7-6 shows the average number 







I \ E 

= 3BeV 

£-1.1 B. 

iV \ 



2 4 6 8 

Thickness of lead, cm 

Fig. 7-6 Development of showers produced by electrons of 1.1 and 3 BeV in 
lead, according to the theory of Snyder. The number of shower electrons is 
plotted against the lead thickness. 



Cosmic Rays 

of shower electrons to be found under various thicknesses of lead 
when electrons of 1.1 BeV and 3.0 BeV energy have initiated the 
showers. These curves bear a striking similarity to the experi- 
mental shower curve in Fig. 7-1. In both the experimental and 
theoretical curves there is an initial fast rise toward a maximum 
followed by a fast drop beyond the maximum; and in both cases 
the maximum appears at a lead thickness of the order of 1 cm. 

Of course, the two curves do not represent precisely the same 
thing. The experimental curve gives the number of coincidences, 
recorded by G-M counters out of line, as a function of lead thick- 
ness, whereas the theoretical curves give the number of shower 
particles emerging from a layer of lead, again as a function of lead 
thickness. Nonetheless, the similarity is certainly not accidental, 
because the probability that a shower emerging from the absorber 
will produce a coincidence increases with the number of particles 
in the shower. 

Detailed analysis of many experimental data showed that on 
the whole the cascade theory satisfactorily explains the observed 
features of showers. Quite illuminating in this regard is the develop- 
ment of showers in materials of different atomic number. As I 
pointed out a little earlier, both the initial rise of the experimental 
shower curve and its subsequent drop are steeper for an element of 
high atomic number such as lead (Z = 82) than for one of low 
atomic number such as aluminum (Z = 13) when the thicknesses 
are measured in grams per square centimeter. Precisely the same 
behavior is found in the theoretical shower curves, the reason being 
that a photon will traverse a smaller thickness of lead than of 
aluminum before undergoing materialization and an electron will 
also traverse a smaller thickness of lead than of aluminum before 
it emits a photon. 

The successful explanation of showers proved the essential 
correctness of Bethe and Heitler's theory and established several 
other important facts: 

1. The local cosmic radiation contains electrons and photons 
with energies of billions of electron volts. 

Electrons, photons, and showers 


2. The observed showers result from cascade processes ini- 
tiated by these electrons and photons. 

3. The individual interactions responsible for the cascades are 
radiative collisions of electrons and pair production by photons. 
These processes occur in the vicinity of atomic nuclei. However, 
they do not involve any change in the structure of nuclei (contrary 
to the earlier belief that showers are the result of nuclear dis- 
integrations). Each interaction gives rise to only two particles (two 
electrons or one photon and one electron). The groups of many 
particles that occasionally appear to diverge from a single point 
arise from several individual interactions occurring next to one 
another in matter. 

4. The ionizing particles that constitute the soft component of 
the local cosmic radiation are the electrons of showers originating 
in the atmosphere or in the roof of the building where the experi- 
ments are performed. 

These results, of course, were very gratifying. However, it was 
still necessary to account for the apparent contradiction between 
the experimental data on showers, which seemed to confirm the 
theory of Bethe and Heitler, and the data on penetrating particles, 
which seemed to disprove it. 

Mu mesons 


The theory of radiation losses and pair production, along with the 
consequent explanation of showers, shifted the mystery from one 
component of the local cosmic radiation to another. To recapitu- 
late, before the theory had made possible any precise predictions 
of the behavior of high-energy electrons and photons, physicists 
had thought the penetrating group of particles (Fig. 6-1) were 
electrons, which in traversing matter lost energy mainly by ioniza- 
tion. Since they were unaware of the overwhelming predominance 
of radiation processes at high energies, they had only to postulate 
electron energies of a few BeV in order to explain the tremendous 
penetrating power of these particles. 

The great puzzle, then, was the nature of the so-called shower- 
producing radiation. Obviously, the production of showers involved 
something more than ionization losses by electrons and Compton 
collisions by photons. After learning about pair production and 
radiation processes, physicists realized that the shower-producing 
radiation consisted of high-energy electrons and photons behaving 
exactly as they ought to behave. But, at the same time, it was 
equally evident that the penetrating particles could not possibly 
be electrons behaving in accordance with theoretical predictions. 



Cosmic Rays 

Mu mesons 


What, then, were these particles? Two obvious possibilities 
suggested themselves, and each had its proponents among cosmic- 
ray physicists. Either the particles are not electrons or, if they are 
electrons, the theory breaks down at energies in the range of several 
BeV. Beyond a certain critical energy, in other words, electrons 
must lose their ability to radiate photons if they are to behave as 
penetrating particles. 

Eventually the second assumption became untenable. On the 
one hand, the British physicist E. J. Williams in 1933 and the 
German physicist C. von Weiszacker in 1935 had shown, on theo- 
retical grounds, that Bethe and Heitler's predictions concerning 
radiation by electrons were bound to be essentially correct up to 
energies well beyond a few BeV. On the other hand, even dis- 
regarding these theoretical arguments, it was difficult to choose a 
critical energy that would explain all the experimental findings. 
Some of the observed showers were so large as to require primary- 
particle energies of many BeV, whereas a large number of the 
penetrating particles appeared to have energies of 1 BeV or less. 
Thus, physicists began to speculate in earnest about the first 

According to theory, if a particle does not radiate as much as 
an electron, it must have a greater mass (Chap. 7). Since the only 
known particle heavier than the electron was the proton, Williams 
suggested in 1934 that the penetrating cosmic-ray particles were 
protons. Williams was also forced to include negative as well as 
positive protons in his hypothesis, for cloud-chamber experiments 
had already shown the penetrating particles to carry ' electric 
charges of both signs. 1 

A new particle of intermediate mass 

The proton hypothesis immediately ran into considerable diffi- 
culties. Indeed, Williams had hardly made his proposal before 

1 Negative protons had been predicted by Dirac's theory as the anti- 
particles of ordinary protons (Chap. 10). 

extensive cloud-chamber studies by Anderson and Seth H. Ned- 
dermeyer at the California Institute of Technology began to un- 
cover evidence directly contradicting it. 

Across their cloud chamber Anderson and Neddcrmeyer had 
placed a lead plate, 3.5 mm thick, for the purpose of studying the 
energy lost by cosmic rays in matter. Many of their pictures 
revealed particles with magnetic rigidities between 10 5 and 10 8 
gauss -cm that did not seem to ionize as heavily as protons of the 
same rigidities should have (Fig. 6-2), but which on traversing the 
plate lost less energy than the radiation theory predicted for 

However, energy loss by radiation is a statistical effect. While 
the theory predicts, on the average, a large energy loss in 3.5 mm 
of lead, it would have been quite possible for some electrons to 
traverse the absorber without passing near an atomic nucleus at 
sufficiently close range to undergo an appreciable energy loss. Thus 
Anderson and Neddermeyer's experiments were very suggestive 
but not entirely conclusive. To obtain more definite evidence, they 
replaced the lead plate with a platinum plate 1 cm thick, which was 
equivalent to nearly 2 cm of lead as far as radiation losses were 
concerned. The probability that an electron would pass through 
such a thick plate without suffering a radiative collision was 
negligible. The results of their new experiments, which they pub- 
lished in the spring of 1937, established the following facts: 

1. There exist two sharply separated groups of particles. The 
first group consists of penetrating particles, which lose a small 
fraction of their energy in traversing a lead plate. The second 
group consists of absorbable particles, which lose a large fraction 
of their energy. The energy loss of the absorbable particles is in 
agreement with the values computed for electrons that radiate and 
ionize according to theory. Hence, the absorbable particles are 
probably electrons. The energy loss of the penetrating particles, 
however, is accounted for entirely by ionization; therefore, the 
penetrating particles do not radiate appreciably. 


Cosmic Rays 

2. Absorbable particles, because they often occur in groups, 
are presumably secondary particles of showers originating above 
the chamber. The particles themselves frequently produce showers, 
confirming the view that they are electrons. Penetrating particles, 
however, usually appear as single tracks in cloud-chamber pictures. 

3. Of the particles whose cloud-chamber tracks show the same 
curvature, some lose large amounts of energy and others lose small 
amounts of energy in traversing the plate. Consequently, electrons 
cannot behave as absorbable particles below a certain critical 
energy and as penetrating particles above that energy. For if they 
did so behave, the energy loss would always be the same for par- 
ticles of the same magnetic rigidity (that is, curvature of tra- 
jectory). This confirms the conclusion that the behavior of pene- 
trating particles cannot be explained by a breakdown of Bethe and 
Heitler's theory, and it proves that the penetrating particles are 
not electrons. 

4. Many of the penetrating particles have magnetic rigidities 
of less than 1.5 X 10 8 gauss -cm. Below this value protons ionize 
at least three times more heavily than fast electrons (Fig. 6-2). 
The ionization density along the tracks of these particles, however, 
is about the same as that of electrons. Since the penetrating par- 
ticles ionize less than protons, their mass must be smaller than the 
proton mass; since they do not radiate as much as electrons, their 
mass must be larger than the electron mass. 

From these results, Neddermeyer and Anderson concluded 
that, in all likelihood, "there exist particles of unit charge, but with 
a mass (which may not have a unique value) larger than that of the 
normal free electron and much smaller than that of a proton." 

The mu meson and its mass 

Neddermeyer and Anderson were unable to determine the mass of 
the new particles with any reasonable accuracy. Nor, for that 
matter, could they rule out the possibility that penetrating par- 

Mu mesons 


tides had the same mass as electrons but possessed some unknown 
property that prevented them from radiating as normal electrons 
radiate. The reason for this uncertainty is clear from the curves in 
Fig. 8-1, where ionization density is plotted as a function of mag- 
netic rigidity for singly charged particles with masses intermediate 
between those of the electron and the proton. At magnetic rigidities 
of about 10 s gauss cm, no charged particle with a mass up to 
several hundred times that of the electron can be distinguished 
from an electron on the basis of the ion trails the two particles 
leave. Only particles whose mass approximates or exceeds the 
proton mass ionize heavily enough to announce their separate 
identity. 1 

From Fig. 8-1 it is also evident that in order to measure their 
mass, it was necessary to observe penetrating particles of fairly low 
energy. This is what J. C. Street and E. C. Stevenson, working at 
Harvard University, had set out to do while Neddermeyer and 
Anderson were carrying out their own experiments. 

Like the latter, Street and Stevenson used a cloud chamber in 
a magnetic field. Since low-energy particles were rare in the pene- 
trating group, they tried to increase the yield of useful pictures by 
observing only particles that stopped in the chamber. To this end, 
they triggered the chamber with an arrangement of G-M counters 
that signaled when a particle entering the chamber failed to come 
out of its walls (Fig. 8-2). They also took pains to operate the 
chamber in such a way as to facilitate the counting of droplets and 
therefore the measurement of the ionization. With this device, 
Street and Stevenson obtained a number of pictures, among which 
was one of particular importance that was published in the fall of 
1937 (Fig. 8-3). The concentration of droplets along the track 
indicates an ionization density of about six times the minimum, 
and the curvature of the track indicates a magnetic rigidity of 

1 Cloud-chamber measurements of ionization densities, which involve 
counting the droplets along a track, are by their nature imprecise. Small 
changes in the adjustment of the chamber cause the vapor to condense on ions 
of one sign alone or, in varying degrees, on ions of both signs. 


Cosmic Rays 

? § I I I § §§!!§ 



Fig. 8-1 Ionization density as a function of magnetic rigidity for singly 
charged particles of different masses. The circle represents the result of density 
and rigidity measurements of the particle track in Fig. 8-3. 

Mu mesons 107 

Determining the mass of a particle from ionization density 
and magnetic rigidity 

The density of droplets along the cloud-chamber track of a particle gives the ionization density. 
If a magnetic field is applied to the chamber, the resulting curvature of the track gives the 
magnetic rigidity. With these two quantities, one can determine the mass of a charged particle 
from the graphs in Fig. 8-1. Here the magnetic rigidity BR is plotted along the horizontal axis 
on a logarithmic scale; the observed ionization density / divided by that of a minimum-ionizing 
particle /„ is plotted on the vertical axis. The various curves represent particles of unit charge 
and different masses; the latter are indicated as multiples of the electron mass along the top 
of the graph. (The particle of mass 1,836 is the proton; the corresponding curve is identical 
with that appearing in Fig. 6-2.) 

The values of the magnetic rigidity and ionization density of a given track define a point 
on the graph. The curve passing through this point indicates the mass of the particle. To take a 
specific example, for the particle track in Fig. 8-3, BR = 9.6 X 10 4 gauss -cm and /// = 6. 
The corresponding point lies close to the curve of 200 electron masses. 

As the graph shows, the curves are widely separated when the magnetic rigidity is not 
too great but are close together when the rigidity increases. At BR = 10* gauss • cm, for 
example, l/l = 7 for a particle of mass 200; and for a particle of mass 300, l/l„ = 13. A crude 
estimate of the ionization is therefore sufficient to distinguish the two particles. At BR = 10* 
gauss -cm, however, both particles have nearly the same (minimum) ionization and cannot be 

9.6 X 10 4 gauss cm. The point corresponding to these two values 
of ionization and rigidity falls close to the "200" curve in Fig. 8-1. 
The particle in question, then, had a mass of approximately 200 
electron masses. 

Thus the work of Street and Stevenson, in addition to estab- 
lishing beyond any doubt the existence of a new particle, also 
provided a fairly accurate estimate of its mass. This new particle 
was variously called baryon, yukon, mesotron, and meson. The last 
name eventually won general acceptance. Later, when investiga- 
tions revealed the existence of other mesons, physicists gave the 
original meson the identifying prefix mu (/i). 

Over the years many physicists endeavored to make precise 
measurements of the mass of y. mesons. Some of the most accurate 
determinations were those of Robert B. Brode and his collaborators 


Cosmic Rays 

G> G, 





Lead obsorber 

) G 3 

Fig. 8-2 Experimental arrangement used by Street 
and Stevenson to obtain cloud-chamber pictures of 
low-energy penetrating particles. The cloud chamber 
C is triggered by particles that discharge the G-M 
counters & to Gj without discharging any of the 
counters G». These particles have sufficient energy 
to traverse the three upper counters and the lead 
absorber but not enough to traverse the walls of the 
cloud chamber and the plate across the chamber. 

O Q Q0. 

at the University of California in Berkeley. Instead of measuring 
magnetic rigidity and ionization, Brode's group measured magnetic 
rigidity and range. The curves in Fig. 8-4 show the relationship 
between these two quantities for particles of different masses and 
illustrate the principle of the method. The arrangement of one of 
the experiments is shown in Fig. 8-5. It consisted of two cloud 
chambers placed one above the other. The upper chamber meas- 
ured the magnetic rigidity ; the lower chamber contained a number 
of lead plates and measured the range. 

The final value of the //-meson mass announced by Brode's 
group in 1950 was 206 ± 2 electron masses. By 1950, y. mesons 
were also being produced artificially in the laboratory by means 
of high-energy particle accelerators. Today the accepted value of 
the /i-meson mass, which comes from a variety of experiments on 
artificially produced mesons, is 206.8 electron masses. 

Mu mesons 


Fig. 8-3 Mass of the *» meson was first determined by Street and Stevenson 
from this cloud-chamber photograph of a /i-meson track (see text). (From a 
paper in The Phytieal Review, vol. 52, p. 1003, 1937.) 

The discovery of y. mesons in 1937 brought to a close the 
second chapter in the history of cosmic rays, which had opened 
with the experiments of Bothe and Kohlhcirster. The nature of the 
local cosmic radiation observed in the atmosphere was now clear: 
The penetrating particles were y. mesons; the absorbable particles 

BR, gouss-cm 

Fig. 8-4 Range in lead as a function of magnetic rigidity for singly charged 
particles of different masses. The vertical bar represents one of the results 
obtained with the experimental arrangement shown in Fig. 8-5. 

were electrons. The nonionizing particles were photons. Mu mesons, 
electrons, and photons accounted for practically all of the radiation 
found near sea level. 

Yukawa's theory 

The discovery of the n meson, like that of the positive electron, 
had been preceded by a theoretical prediction. In 1935 the Japanese 

Mu mesons HI 

Determining the mass of a particle from magnetic rigidity and range 

The graphs in Fig. 8-4 can be used to determine the mass of a particle whose magnetic rigidity 
and range have been measured experimentally. The magnetic rigidity BR is plotted along the 
horizontal axis on a logarithmic scale; the range, in grams per square centimeter, is plotted 
along the vertical axis, also on a logarithmic scale. The various curves represent particles of 
unit charge and different masses (in multiples of the electron mass). The magnetic rigidity 
and the range of a single particle define a point on the graph. The curve passing through this 
point determines the mass of the particle. 

In the experiments of Brode (see Fig. 8-5) the measurements of range in lead contained 
an uncertainty of 0.63 cm, or about 7 g/cm'. The vertical bar in Fig. 8-4 represents a typical 
case, in which the range of the particle falls somewhere between 44 and 51 g/cm« and the 
magnetic curvature is 5 X 10 s gauss -cm. From the graph one can see that the mass of the 
particle lies between 170 and 200 electron masses. The uncertainty in the measurement of 
magnetic rigidity makes this mass determination consistent with the assumption that the 
particle is a n meson. The over-all accuracy obtained by averaging a large number of measure- 
ments is, of course, considerably greater than that of any individual measurement. 

Fig. 8-5 Schematic diagram of the arrange- 
ment used by Brode and his collaborators to 
measure the mass of it mesons. Both cloud 
chambers & and & are triggered by particles 
that discharge the three G-M counters & to G». 
Chamber & is placed in a magnetic Geld; the 
curvature of the tracks seen in this chamber 
gives the magnetic rigidity. Chamber C» con- 
tains 15 lead plates, each 0.63 cm thick. When a 
particle stops in this chamber, its range can be 
determined with an uncertainty equal to the 
thickness of a single plate. 


Cosmic Rays 

physicist Hideki Yukawa had postulated the existence of a sub- 
atomic particle several hundred times heavier than the electron. 
I cannot do justice to Yukawa's theoretical reasoning within the 
bounds of this book, and what follows is hardly more than a crude 
attempt to lend some small degree of plausibility to his con- 

As I mentioned briefly in Chap. 2, nature has obliged us to 
accept the fact that all radiations behave in certain respects as 
waves, in other respects as particles. The waves associated with 
photons are electromagnetic waves. As such they are related to the 
electric forces between charged particles. In fact, quantum theory 
reveals a close relationship between the properties of photons and 
the nature of electric forces. 

Electric forces of attraction between the positive nucleus and 
the negative electrons of an atom are the forces that hold the atom 
together. By 1935 the nucleus itself was known to contain both 
protons and neutrons. 1 Now, protons repel each other because of 
their electric charges, whereas neutrons do not interact electrically 
with one another or with protons. Yet atomic nuclei are remarkably 
stable structures. This fact had made it necessary to assume forces 
of attraction of a nonelectric nature between protons and neutrons. 
Over the small distances characteristic of atomic nuclei these 
nuclear forces are much stronger than electric forces; over greater 
distances they are much weaker. In other words, nuclear forces 
decrease with distance far more rapidly than they would if they 
behaved according to the inverse-square law, which governs 
electric forces. 

Yukawa argued that, just as electric forces are associated with 
the photon, so nuclear forces should be associated with some kind 
of particle. He then proceeded to study the properties of this 
hypothetical particle and found that the short range of nuclear 
forces required a particle of finite mass; in fact, a mass of the order 
of several hundred electron masses. Yukawa also predicted that 
the particles associated with the nuclear field would be unstable; 

1 See Appendix G. 

Mu mesons 


that is, would decay much as the nuclei of radioactive substances 
do. Furthermore, the mean life 1 of the u meson, before it decays, 
should be of the order of 1 microsecond (10 -6 second), and the dis- 
integration products of each particle should be one electron and one 
neutrino. The latter was a particle with zero mass and no electric 
charge whose existence had been postulated by Wolfgang Pauli 
and Enrico Fermi in order to explain why the electrons emitted in 
the decay of a given radioactive element have a variety of 
energies. 2 

After the discovery of mesons, some physicists, recalling 
Yukawa's theory, concluded that the penetrating particles in 
cosmic radiation were the "quanta of the nuclear force field" pre- 
dicted by the theory — hence the name yukon. As will be seen in 
Chap. 9, this identification proved incorrect. Moreover, when the 
quantum of the nuclear force field was found, its properties differed 
somewhat from those envisaged by Yukawa. Nevertheless, the 
idea that cosmic-ray mesons might be Yukawa particles led to a 
very important discovery. 

The discovery of mu-meson decay 

Between 1936 and 1937, scientific teams in England, France, Ger- 
many, and Italy had made accurate measurements of the numbers 
of cosmic-ray particles found at various altitudes in the atmos- 
phere. And the results had been rather puzzling. Contrary to the 

1 The concept of mean life is taken over from the theory of radioactivity, 
according to which it represents the average life span of all nuclei in a given 
radioactive sample. The mean life happens to coincide with the period in 
which the total number of nuclei is reduced, by decay, to 1/2.7 of its original 
value. (The number 2.7 is the base of natural logarithms; the definition of 
mean life is, from a mathematical point of view, similar to the definition of 
mean free path; See page 24 and Fig. 2-4.) The reader is perhaps more familiar 
with the concept of half-life, the period in which the total number of nuclei is 
reduced to one-half its original value. The mean life of a radioactive nucleus 
is 1.4 times its half-life. 

1 See Appendix H. 


Cosmic Rays 

earlier findings of Millikan's group (Chap. 1), it looked as if air 
absorbed cosmic rays more effectively than solid or liquid matter 
did when layers of the same mass per unit area were compared. 
Moreover, the thinner air at very high altitudes appeared to be a 
better absorber than the denser air in the lower atmosphere. 

Now, Yukawa particles were supposed to be unstable. If 
cosmic-ray mesons were identical with Yukawa particles, they 
should decay spontaneously. In 1938 the German physicist H. 
Kuhlenkampff pointed out that this property offered a natural 
explanation for the anomalous absorption of cosmic-ray particles 
in air. The argument is very simple (Fig. 8-6). Consider, for 
example, a layer of water 10 cm thick. Near sea level a layer of air 
of the same mass per unit area is 8,000 cm thick, since air is 800 
times less dense than water. At an altitude of about 5,100 meters, 
where the air density is one-half that at sea level, the equivalent 
air layer is 16,000 cm thick. A particle moving at nearly the speed 
of light will travel 10 cm in 3.3 X 10 _, ° second, or 0.00033 micro- 
second. The same particle will need 0.265 microsecond to travel 
8,000 cm and 0.53 microsecond to travel 16,000 cm. 

Suppose now that mesons have a mean life of the order of 
microseconds. Then practically none of them will decay while 
traversing 10 cm of water. Only mesons with a range less than 10 
cm of water will be stopped by this absorber. Many mesons, how- 
ever, will undergo spontaneous decay while traversing 8,000 cm of 
air, and even more will do so while traversing 16,000 cm of air. 
Consequently, to the normal absorption of mesons is added the 
effect of spontaneous decay. As a result, air will appear to be a 
better absorber than condensed matter, the more so the thinner 
it is, which is exactly what the experiments had indicated. 

Kuhlenkampff's idea was taken up by a number of investi- 
gators, including Werner Heisenberg and H. Euler in Germany and 
Blackett in England, who worked out its various consequences. 
Indeed, from late in 1938 to the end of 1939 the question of the 
radioactive decay of mesons was one of the most hotly debated 
subjects in cosmic-ray physics. The experimental evidence from 

M u mesons 




Fig. 8-6 Comparison of meson absorption in layers of water and air of the 
same mass per unit area. Equal numbers of mesons come to rest in the two 
layers. However, some mesons undergo spontaneous decay in traveling the 
greater distance through the air layer. Thus, fewer mesons are found under the 
layer of air than under the equivalent layer of water. 


Cosmic Bays 

atmospheric absorption was highly suggestive, but not compelling. 
In addition, attempts to detect the decay of mesons directly in 
cloud-chamber experiments had all failed. 

I became interested in this problem myself while visiting the 
Institute for Theoretical Physics in Copenhagen, directed by Niels 
Bohr, in the fall of 1938 and Blackett's laboratory at the University 
of Manchester during the following winter. In the spring of 1939 I 
moved to the University of Chicago at the invitation of A. H. 
Compton. At the time, I was toying with the idea of testing the 
hypothesis of the instability of mesons by making a direct and 
precise comparison of meson absorption in air and in some dense 

Compton most generously put at my disposal the means for 
carrying out this experiment. Two members of his group, Norman 
Hilbcrry and J. Barton Hoag, offered their assistance. We set up, 
in a truck, a meson detector consisting of three G-M tubes in 
coincidence, with enough lead between and around the counters to 
cut off the electron component (Fig. 8-7). With this equipment we 
took measurements in Chicago (180 meters above sea level) ; then 
we drove to Colorado and took measurements at Denver (1,600 
meters), Echo Lake (3,240 meters), and Mt. Evans (4,300 meters). 
At each of the Colorado stations we counted mesons both with and 
without a layer of graphite (about 87 g/cm 2 ) above the counters. 

The results are presented in Fig. 8-8. They showed the absorp- 
tion in graphite to be much less than that in air. For example, 
starting from an atmospheric depth of 699 g/cm 2 (Echo Lake), 
the addition of 87 g/cm 2 of graphite cut down the number of 
mesons by 10 per cent, while the addition of 87 g/cm 2 of air 
reduced this number by 20 per cent. The additional apparent 
absorption was due to the decay of mesons "in flight" through the 

From the measurement of the additional absorption we were 
able to estimate roughly the mean life of mesons. Tn making this 
estimate, we had to take into account the fact that, according to 
the theory of relativity, the determination of a time interval de- 

Mu mesons 


pends on the frame of reference in which the measurement is made. 
If, for example, a clock is moving with a velocity v with respect to 
some observer, it will appear to tick at a slower rate to that observer 
than to another observer moving with the clock. The time intervals 
between ticks, as measured by the f irst and s econd observer, 
respectively, are in the ratio of \/Vl - v 2 /c 2 , where c is the 
velocity of light. 

When different values are substituted for », it is clear that the 
effect becomes appreciable only when the velocity of the clo ck is very 
close to c (that is, when the value of the denominator Vl - v 2 /c 2 
approaches zero). Now, unstable cosmic-ray mesons are subatomic 
"clocks" moving at very great speeds. The measured mean life 
t' of these mesons is therefore markedly greater than the mean 
life t of mesons at rest. For example, it is found that for mesons 
with a kinetic energy of 500 MeV, t'/t = 4.7. In other words, the 
measured mean life of these mesons is 4.7 times longer than their 

Fig. 8-7 Experimental arrangement for the com- 
parison of the meson absorption in air and graphite. 


Cosmic Rays 

mean life at rest. 1 Therefore, to determine the mean life of mesons 
at rest from measurements involving mesons in motion, it was 
essential to know their energy. By making suitable estimates of the 
energy (or, more precisely, of the average energy of the mesons we 
were observing) we found a mean life for mesons at rest of the order 
of 2 microseconds (2 X 10 -0 second). 

1 From the relativistic expression of kinetic energy E [see Appendix D, 
Eq. (D-2)l, it follows that 1/Vl - oVe 3 = (£'/107) + 1. In this equation IS is 
measured in MoV and 107 represents the "rest energy" of the ii meson, m&, 
also in MeV. 


$ 8 

V s * 
X ^ 
\ N. 

X x. 

\ ^ 

X s * 
X v. 

X *•* 


800 900 

Total depth, g/cm 2 



Fig. 8-8 Results of meson-absorption measurements in air and graphite 
(made with the arrangement shown in Fig. 8-7) . Open dots are counting rates 
per minute observed at Chicago, Denver, Echo Lake, and Mt. Evans with no 
absorber above the counters. Solid dots are counting rates observed at the 
three higher stations under 87 g/cm 8 of graphite. The horizontal scale gives 
the total thickness of matter (air plus graphite) above the counters, measured 
in grams per square centimeter. 

Mu mesons 


Similar experiments were performed in the following years by 
many physicists, including W. M. Nielsen, H. Victor Neher, 
Martin A. Pomerantz, and their collaborators in the United States 
and Gilberto Bernardini, Giuseppe Cocconi, Oreste Piccioni, and 
others in Italy. On the whole, their results were in agreement with 
ours. In passing, I should like to recall one such experiment, per- 
formed by David B. Hall and myself, in which we measured the 
apparent mean lives of mesons of different energies and found that 
they changed with energy just as the theory of relativity predicted. 
The effect was quite large — about a factor of 3 from the fastest 
to the slowest mesons detected in the experiment. 

Electrons from the decay of mu mesons 

In the meantime, Williams and G. E. Roberts in England had 
succeeded in obtaining the first cloud-chamber picture of meson 
decay. The meson in question, positively charged, had stopped in 
the gas of the chamber (a very rare event indeed), and out of the 
end of its track had appeared the track of a positive electron. This 
picture, besides providing final, crucial proof of meson decay, also 
showed that (as anticipated) when a ti meson disappears, an 
electron is born. (Many pictures of this kind were obtained in later 
years. Figure 8-9 shows a particularly clear example from experi- 
ments performed in 1948 by Robert W. Thompson, then a graduate 
student at the Massachusetts Institute of Technology.) 

Direct measurements of the mean life 
of mu mesons 

The next major step was taken by Franco Rasetti. Williams and 
Roberts' photograph had shown pictorially the decay of a meson 
brought to rest by ionization losses in matter. Mesons stopping in 
matter should not decay immediately, but should "wait around" 
for a period equal, on the average, to their mean life. Since the 

Fig. 8-9 Decay of a M meson. The meson enters the cloud chamber from 
above. It traverses an aluminum plate 0.63 cm thick, where it loses most of 
its energy. The meson, which leaves the plate as a slow and therefore heavily 
ionizing particle, comes to rest in the gas. The track of an electron originates 
from the end of the M-meson track. The electron, traveling at nearly the speed 
of light, produces a track approximating that of a minimum-ionizing particle. 
The tracks of the meson and the electron are slightly bent by a magnetic field, 
and the direction of the deflection shows that both particles are positively 
charged. (From R. W. Thompson, The Physical Review, vol. 74, p. 490, 1948.) 

Mu mesons 


mean life had already been estimated to be several microseconds, 
it was expected that electrons arising from the decay of m mesons 
would come out of an absorber after an average delay of a few 

In an attempt to demonstrate directly the delayed emission of 
electrons by mesons brought to rest in matter, Rasetti in 1941 
performed the experiment illustrated schematically in Fig. 8-10. 



) 6 

Circuit selecting 

stopped mesons 


Time-delay o- 
o coincidence o- 
circoits q. 




Fig. 8-10 Decay of m mesons at rest was observed in 1941 by Rasetti, using 
the experimental apparatus shown here. A m meson discharges counters 
Gi to G« and then stops in the iron absorber. (A particle passing through the 
absorber would discharge one of the counters, G,.) A short time later, its decay 
electron (if it comes out of the absorber in the right direction) discharges one 
of the side counters Gt. Three different electronic circuits select events in 
which the discharge of counters G t occurs within 1, 2, or 15 microseconds (jis), 
respectively, after the discharge of counters G, to G t . 


Cosmic Rays 

The G-M counters G, to G 4 and G. and the associated electronic 
circuit selected mesons that stopped in the iron-block absorber. 
These mesons discharged one of the three G, counters as well as the 
single counters G 2 , G,, and G 4 , but they did not discharge any of the 
G 6 counters. Along each side of the absorber were six counters G 6 
whose function was to detect the electrons arising from the decay 
of stopped mesons. A coincidence circuit gave a signal whenever 
one of the G 6 counters was discharged less than 1 microsecond after 
the arrival of a meson. Two other coincidence circuits, operating 
in the same manner, had allowed delays of 2 and 15 microseconds. 
Rasetti found that the 2-microsecond circuit recorded more 
coincidences than the 1-microsecond circuit. This meant there were 
electrons being emitted with a delay greater than 1 microsecond 
but less than 2. Rasetti also found that the 15-microsecond circuit 
recorded more coincidences than the 2-microsecond circuit. Thus, 
a considerable fraction of the electrons were being emitted after a 
delay of more than 2 microseconds but less than 15. As expected 
from previous estimates, then, the electrons arising from the decay 
of stopped mesons had come out of the absorber over a time 
interval of the order of microseconds. On the basis of the relative 
numbers of coincidences recorded by the three different circuits, 
Rasetti estimated the mean life of mesons at rest to be about 1.5 

In the meantime, I had moved to Cornell University, where I 
had started a research program in cosmic rays together with two 
young associates, Kenneth I. Greiser and Norris G. Nereson. In 
1943 Nereson and I made a precise study of meson decay by using 
an arrangement of G-M counters and absorbers similar to Rasetti's. 
The novel feature of our experiment was an "electronic clock" that 
we built for the purpose and that enabled us to measure the indi- 
vidual delays between the arrival of mesons and the emission of 
electrons. We found that the decay of mesons, like the decay of 
radioactive atoms, is exponential (Fig. 8-11). This means that, out 
of any large group of mesons present at a given time, a fixed frac- 
tion will decay during the next microsecond regardless of how long 

1.SX10 4 

4 3 6 


Fig. 8-11 Decay curve of m mesons brought to rest in lead, as measured by 
Nereson and the author. Shown on the horizontal axis is the time interval 
following the arrival of a meson in the absorber; on the vertical axis is shown 
the number of observed n mesons whose decay occurred after the corresponding 
interval indicated on the horizontal axis. For example, 10,000 mesons survived 
more than 1 microsecond, 6,300 mesons survived more than 2 microseconds, 
and so on. 


Cosmic Hays 

the individual mesons have already "lived." From our measure- 
ments, we obtained for the mean life a value of 2.15 microseconds, 
with an estimated error of 0.1 microsecond. 

In the next few years several other physicists made similar 
measurements on cosmic-ray mesons and confirmed the value we 
had found. Among them were R. Chaminade, A. Freon, and R. 
Maze in France and Marcello Conversi and Piccioni in Italy. 1 
After it became possible to produce ^ mesons with high-energy 
accelerators, their mean life was measured with greater accuracy 
and turned out to be 2.212 microseconds. 

1 To appreciate the contributions of the European scientists, one must 
consider the severe handicaps under which they worked in the war years. The 
Italians, for example, did most of their experiments while hiding in a cellar, 
where they had smuggled their equipment when the Germans descended on 
Rome in 1943. 

Pi mesons 


The history of scientific exploration, like the history of geographical 
exploration, is full of instances in which an unexpected discovery 
overshadows the original goal. Columbus, searching for a new 
route to India in the last decade of the fifteenth century, discovered 
America. Physicists, searching for a solution to the cosmic-ray 
puzzle in the fourth decade of the twentieth century, came across 
two previously unknown particles, the positron and the n meson. 
The discovery of these particles, which arise from the interactions 
of high-energy rays with matter but are not among the building 
blocks of ordinary matter, opened an entirely new field of research. 
Here cosmic-ray physicists found such a wealth of novel and 
unexpected phenomena that for many years the exploration of this 
field was their main concern. 

Eventually the development of accelerators in the multi- 
billion eV range, by providing controlled and much stronger 
sources of high-energy particles, brought a decline in cosmic-ray 
research as a means for studying elementary particles. Before that 
happened, however, cosmic-ray physicists had succeeded in draw- 



Cosmic Hays 

ing a fairly comprehensive, though not entirely complete or very 
detailed, map of the entire field of elementary-particle physics. 

The puzzling behavior of negative mu mesons 

For physicists interested in the atomic nucleus, as well as for those 
interested in cosmic rays, the most important development in the 
exploration of the new world of particle physics came with the 
discovery of the pi meson (a-) in 1947. The investigations leading 
to this discovery began in 1940, when the Japanese physicists 
Sin-itiro Tomonaga and Gentaro Araki pointed out that positive 
and negative » mesons should behave in a characteristically differ- 
ent manner after coming to rest in matter. 

A positive meson, they argued, can never approach very near 
an atomic nucleus because of the electric forces of repulsion exerted 
by the positive nuclear charge. Therefore, it will "wait around" 
until it disappears by spontaneous decay. A negative meson, how- 
ever, is attracted by the nuclear charge and thus might be captured 
by a nucleus before it has a chance to decay. If the interaction 
between M mesons and the nuclei of an absorber is sufficiently 
strong, almost all the negative » mesons will be swallowed up by 
nuclei; few of them will decay into electrons. If the interaction is 
weaker, some will be absorbed and some will decay. 

In any case, the life expectancy of negative mesons must be 
less than that of positive mesons, for the latter die a "natural" 
death, while the former are exposed to the additional risk of "acci- 
dental" death by nuclear capture. Once captured, negative mesons 
presumably disappear, and their mass turns into energy according 
to Einstein's equation. Like a tiny bomb, such a release of energy 
should cause the nucleus to explode. 

Shortly after Tomonaga and Araki published their paper, 
Rasetti performed the experiment on /i-meson decay described in 
the preceding chapter. By comparing the number of mesons stop- 
ping in his iron block with the number of decay electrons coming 

Pi mesons 


out of it, Rasetti was able to show that only half the mesons de- 
cayed into electrons. He concluded that the decay electrons came 
from positive mesons and that the negative mesons disappeared 
by nuclear capture, as Tomonaga and Araki had predicted. 

A more direct experimental proof developed from the investi- 
gations of Conversi, Pancini, and Piccioni. The experiment they 
performed was similar to the earlier experiments by Rasetti and by 
Nereson and myself, except for the addition of a device capable of 
concentrating upon the absorber either positive or negative mesons. 
This device, whose design goes back to my own early work on 
cosmic rays in 1931,' is known as a magnetic lens. It takes advan- 
tage of the strong magnetic fields that can be generated in iron and 
also of the fact that cosmic-ray mesons can traverse fairly large 
thicknesses of iron without suffering much energy loss or scattering 
(which is not true of particles of lower energy, such as a or rays 
from radioactive sources). Thus one can use magnetized iron to 
defied cosmic-ray mesons. Since it is much easier to magnetize a 
piece of iron than to produce an equivalent field in air, the device 
also has great practical advantages. 

The principle of the magnetic lens is illustrated in Fig. 9-1. 
Two bars of iron placed side by side are magnetized in opposite 
directions. Two G-M counters, one above and one below the bars, 
are connected in coincidence. Suppose the magnetic field of the bar 
on the right points toward the observer and the field of the bar on 
the left away from the observer. Positive mesons traveling down- 
ward through the first counter will be deflected toward the second 
counter, thereby increasing the number that passes through both. 
Negative mesons (except for the very few that travel straight down 
between the bars) will be deflected away from the second counter. 

' At that time, of course. I did not know about mesons. I had developed 
the lens in an attempt to find out whether the cosmic-ray particles capable of 
traversing largo thicknesses of matter were positively or negatively charged. 
The result of the experiment was ambiguous; I could only conclude either that 
the particles had too much energy to undergo appreciable magnetic deflection 
in the magnetb-ed iron of the lens or that positive and negative particles were 
about equally abundant. 


Cosmic Rays 

If the magnetization is reversed in both bars, the lens will concen- 
trate negative mesons and reject positive mesons. 

With an iron absorber below their lens, Conversi and his two 
colleagues found only positive mesons giving rise to the delayed 
electrons characteristic of spontaneous decay. Thus they concluded 
that practically all negative mesons stopping in iron were captured 
by iron nuclei before they had a chance to disintegrate spon- 

Between 1947 and 1948, the Italian group, as well as research 
teams at the University of Chicago and the Massachusetts Insti- 
tute of Technology, extended the decay experiments to a number of 
different substances. The most important result, first announced 
by the Italians, revealed that capture by the atomic nuclei of light 
elements (for example, carbon) did not compete significantly with 
the spontaneous decay of negative mesons. The probability of 


♦ ♦♦ 



• • • 

• • • 

(a) {b) 

Fig. 9-1 Magnetic lens consisting of two iron bars magnetized in opposite 
directions, (a) The magnetic field points away from the observer (that is, into 
the page) in the bar at left and toward the observer (that is, out of the page) in 
the bar to the right. Positive particles passing through the upper G-M counter 
are deflected by the magnetic field toward the lower counter. (6) The mag- 
netization of the bars is reversed. Positive particles passing through the upper 
G-M counter are deflected away from the lower counter. The lens in (a) will 
deflect negative particles away from the lower counter; the lens in (6) will 
deflect them toward the counter. 

Pi mesons 


nuclear capture appeared to increase gradually with increasing 
atomic number. In magnesium (Z = 12) about one-half of the 
negative y. mesons disappeared by decay, one-half by nuclear 
capture. In elements heavier than magnesium, nuclear capture 
exceeded spontaneous decay; in elements lighter than magnesium, 
the reverse was true. 

The fact that nuclear capture became increasingly effective 
with increasing atomic number was not in itself surprising. Heavy 
nuclei have a greater positive charge and therefore must provide a 
more effective trap for negative mesons to fall into. Physicists, 
however, found it extremely difficult to understand how, even in 
elements as light as carbon (Z = 6), negative mesons could possibly 
escape nuclear capture. Their argument ran as follows. Mu mesons 
were known to have a mean life of about 2 microseconds. Now, 
2 microseconds may seem a very short interval, but on the atomic 
scale of time it was very long. In fact, it was at least 20 million 
times longer that the most generous theoretical estimate of the 
interval required for nuclear capture. 

In other words, one could be sure on theoretical grounds that 
all negative y mesons, once they were stopped in carbon, would fall 
into carbon nuclei before they had a chance to decay. If, neverthe- 
less, most mesons did undergo spontaneous decay, they must retain 
their identity within the nuclei for a time interval so great that they 
had a chance to come out again. In the language of modern physics, 
this meant an exceedingly weak interaction between the y mesons 
and the protons and neutrons of carbon nuclei. 

Such a weak interaction was puzzling, to say the least. In the 
first place, if y mesons were the "quanta of the nuclear force field" 
predicted by Yukawa (see Chap. 8), they had to interact strongly 
with protons and neutrons. In the second place, it was thought, 
n mesons were produced in the passage of high-energy particles 
through nuclei, an event that lasts an extremely short time, some- 
thing of the order of lO"" second. If this were the case, why did the 
opposite process, namely, the absorption of y. mesons by nuclei, 
require periods of several microseconds? The answer, when it came, 


Cosmic Rays 

was another instance of the close interplay between technical 
developments and scientific advances, of which the history of 
science offers so many examples. 

Nuclear emulsions 

The cloud chamber is a truly wonderful tool. It has enabled 
physicists to "see" elementary particles and the results of their 
collisions with atoms or nuclei. But, like any experimental device, 
the cloud chamber has inherent limitations. Because of the low 
density of gases, very few of the particles entering a cloud chamber 
collide with nuclei or stop within the chamber. 

To improve the situation, physicists, in the 1930s, began to 
build larger chambers and to place slabs of solid materials (such as 
carbon or various metals) across them. The technique of mulliplale 
chambers proved very useful in a number of studies, some of which I 
shall relate in the next chapter. Still, it was not an entirely satis- 
factory solution to the problem; for when a particle interacts 
within a plate (which is usually from a few millimeters to a few 
centimeters thick), it is impossible to see exactly what happens at 
the point where the collision occurs. Similarly, when a particle 
stops in a plate and then decays, the decay products can be 
observed only if they have enough energy to come out of the plate. 
The direct observation of interactions and decay processes requires 
some dense substance in which particles have a good chance of 
undergoing collisions or of coming to rest and in which they will 
somehow leave visible tracks. 

In the middle 1940s physicists succeeded in perfecting a de- 
tector with the desired properties, the so-called nuclear emulsion. 
Photographic emulsions had, of course, been widely used in radia- 
tion studies for many years. They had been responsible for Boent- 
gen's discovery of X-rays, had played an important role in nuclear 
physics, and had been used occasionally in cosmic-ray research. 

The effect of high-energy radiations upon a photographic 
emulsion is similar to that of light. Ionizing particles "sensitize" 
the grains of silver bromide that they encounter along their path 

Pi mesons 


in the emulsion. An appropriate "developer" solution will then 
reduce the sensitized grains to silver. Under a microscope, the 
trajectories of individual ionizing particles appear as rows of dark 
silver grains. But ordinary emulsions are sensitive only to com- 
paratively slow particles, which leave a very dense trail of ions in 
their wake. Moreover, until the middle 1940s the available emul- 
sions were exceedingly thin, and only particles traveling almost 
exactly parallel to the photographic plate left a track of any 
appreciable length. For these reasons, photographic emulsions had 
been almost completely abandoned as a means of detecting charged 


The emulsion technique was revived by a group of physicists 
at the University of Bristol in England under the leadership of 
C. F. Powell and Occhialini. Working in collaboration with 
scientists at the Ilford Company, they prepared new emulsions 
that contained a much higher concentration of silver bromide than 
ordinary photographic types and were thus more sensitive to 
ionizing particles. Soon afterward, scientists of the Kodak Com- 
pany in the United States were also working in this field and 
making important contributions to its progress. New and ingenious 
methods for developing exposed plates made it possible to use 
emulsions almost a millimeter thick, or more than 100 times 
thicker than those previously available. A later technical develop- 
ment was the stripped emulsion, an emulsion without any glass 
backing, which could be stacked in layers to form what became 
known as emulsion chambers. 

At the same time, physicists became experienced in recognizing 
the properties of charged particles from the appearance of their 
tracks in nuclear emulsions. When the emulsion is developed under 
carefully controlled conditions, the density of the black silver 
grains along a track is proportional to the density of ion pairs that 
the particle would produce in a gas. Thus the density decreases 
with increasing velocity in a known and well-defined manner. 

Another measurable characteristic is scattering, or the "wig- 
gliness" of a track. The scattering decreases as the kinetic energy 
increases, but it also depends to some extent on the mass of the 

Pi mesons 


Measuring the mass of a particle from the grain density along its track 
in a nuclear emulsion 

Figure 9-2 can be used to determine the mass of a particle that stops in a nuclear emulsion. 
Plotted on the horizontal axis is the residual range R, in centimeters, that is, the distance 
of a given point from the end of the track. Plotted on the vertical axis is the grain density g 
around this point, divided by the grain density of a minimum-ionizing particle g„. The various 
curves represent particles of unit charge and different masses; the latter are indicated as 
multiples of the electron mass along the top of the graph. By measuring g at various points 
along a given track, one obtains a series of points; the curve that fits these points most closely 
indicates the mass of the particles. Shown in the figure are experimental points obtained by 
the Bristol group for protons (mass 1,836), for x mesons (mass 273), and for particles of mass 
966 (the K mesons, discussed in chapter 10). 

particle. Finally, if a particle stops in the emulsion, one can 
measure its range; the range, too, depends on the energy and the 
mass of the particle. Consequently, by measuring any two of the 
three quantities grain density, scattering, and range, it becomes 
possible to determine the particle mass (Fig. 9-2). In principle, 
this method for determining mass is similar to that used in cloud- 
chamber experiments (Figs. 8-1 and 8-4). 

The discovery of the pi meson 

In 1947 C. M. G. Lattes, Occhialini, and Powell exposed some of 
the newly developed nuclear plates to cosmic rays at mountain 
altitudes. In these plates they found several tracks of the kind 
shown in Fig. 9-3. The picture shows a particle (*) entering the 
emulsion and coming to rest. The gradual increase of grain density 
due to the gradual decrease in velocity leaves no doubt as to the 
direction in which the particle is traveling. The rate at which the 
grain density increases as the particle approaches the end of its 
range indicates a mass several hundred times that of an electron 
(Fig. 9-2). From the point at which this particle comes to rest, 
another particle (m) emerges. The second particle, which comes to 
rest in the emulsion after traveling a distance of slightly more 

Pi mesons 


than 0.5 mm, also appears to have a mass of a few hundred electron 
masses. However, careful measurements of the grain densities at 
different distances from the end of the range show conclusively 
that the second particle is somewhat lighter than the first. 

After discarding several interpretations, Lattes, Occhialini, 
and Powell came to the conclusion that the picture showed the 
decay of one meson into a lighter meson. But the heavier "parent" 
particle could not be the meson associated with the penetrating 
component of the local cosmic radiation, because when that meson 
decays, it produces an electron, not another meson. Thus, unless 
there were more than two kinds of mesons, it was the lighter, 
"daughter" particle that had to be identified with the previously 
known meson. 

If this explanation were correct, the secondary meson, after 
coming to rest in the emulsion, should decay and produce an 
electron. The emulsions used in the 1947 experiments were not 
sufficiently sensitive to detect minimum-ionizing particles such as 
fast electrons. When emulsions sensitive to those particles became 
available in 1949, the decay electron duly appeared (Fig. 9-4). 
There was no longer any doubt about the conclusion reached by the 
Bristol group. The heavier, parent particle became known as the 
t meson. The lighter secondary particle, which until that time had 
been the only meson, was called the ix meson. 

The discovery of the *■ meson clarified the meson puzzle 
considerably. The "quanta of the nuclear force" predicted by 

Fig. 9-4 Photomicrograph of tracks in a nuclear emulsion, showing a v 
meson (x) that comes to rest and decays into a m meson (ji)- The p. meson in 
turn comes to rest and decays into an electron (e). (From R. H. Brown, 
U. Camerini, P. Fowler, H. Muirhead, C. F. Powell, and D. M. Ritson, 
Nature, vol. 163, p. 47, 1949.) 



Cosmic Rays 

Yukawa were t mesons, not /x mesons. Pi mesons were the par- 
ticles produced in high-energy nuclear interactions. The fact that 
p mesons did not interact strongly with nuclei was no longer sur- 
prising, because n mesons were not produced in nuclear inter- 
actions, but were born through the decay of r mesons. This picture, 
however, was quite different from the earlier, theoretical one. 
Yukawa had thought his particles would decay into electrons. 
Instead, they decayed into it mesons, and it mesons in turn decayed 
into electrons. 

But what about negative ir mesons, the original source of the 
meson puzzle? If tt mesons were actually the particles of Yukawa's 
theory, and if they were abundantly produced in nuclear inter- 
actions, then negative ir mesons that stopped in matter were also 
captured promptly by nuclei. As I have already pointed out, the 
energy suddenly released by the disappearance of a captured meson 
should produce a nuclear explosion. That is exactly what happens. 
Negative ir mesons coming to rest in a nuclear emulsion are never 
seen to decay into it mesons. Rather, a "star" appears at the end 
of their track, as was first shown by D. H. Perkins and by Occhialini 
and Powell in 1947 (Fig. 9-5). The star is produced by the particles 
into which the nucleus disintegrates as a result of the explosion. 
Thus, unlike a negative it meson, a negative t meson that stops in 
matter has no chance to decay before a nucleus commits the 
"suicidal" act of trapping and "killing" it. 1 

Since the mesons found in the cosmic radiation near sea level 
were almost entirely of the ti variety, r mesons had to decay so fast 
that practically none of them survived the journey through the 
atmosphere. The earliest calculations indicated a mean life con- 
siderably shorter than 1 microsecond. This surmise proved to be 

1 It is interesting to note that when x mesons stop in an emulsion, they 
leave an unmistakable signature: a t-ii decay in the case of a positive x meson 
(Figs. 9-3 and 9—4) and a nuclear star in the case of a negative x meson (see 
Fig. 9-5). Thus the nuclear emulsion technique is ideally suited for the study 
of low-energy x mesons, which is the main reason why it has been so widely 
applied in studies of cosmic radiation and in experiments with high-energy 

Pi mesons 

correct even though the argument on which it rested was not 
entirely foolproof (ir mesons fail to travel as far as it mesons 
because, besides decaying more readily, they interact more fre- 
quently with atomic nuclei). Experiments carried out between 1948 
and 1950 with high-energy accelerators gave a value of 2.5d X 10 
second for the mean life of * mesons, or about 100 times shorter 
than the mean life of „ mesons. High-energy accelerators also made 
possible a precise measurement of the ir-meson mass, which was 
found to be 273 electron masses. 

When the mean life and mass of both * and » mesons had been 
determined, there still remained the question of what disintegration 
products arose in the decay of these particles. Yukawa had pre- 

Fig. 9-5 Nuclear capture of negative x meson. The meson comes to rest ma 
nuclear emulsion and is promptly captured by a nucleus. The mass of the 
meson is immediately changed into energy, causing the nucleus to explode 
The star in the picture is composed of the tracks left m the emuls.on by he 
charged fragments of the nucleus. (From a paper by C. F . Powel „ Colston 
Papers. Butterworth & Co. (Publishers), Ltd., London, 1949, p. 83.) 


Cosmic Rays 

dieted that his quantum of the nuclear force field would decay into 
an electron and a neutrino. But the charged particles appearing in 
the decay of t mesons were ft mesons rather than electrons. And the 
v mesons, which did give rise to electrons, had nothing to do with 
Yukawa's theory. Consequently, there was no reliable theory on 
which to base any prediction concerning the decay of either meson. 
Only experiment could provide the answer. 

In the decay of a meson (as in the decay of a radioactive atom) 
a certain amount of mass is transformed into energy in accordance 
with Einstein's equation E = mc*. The mass that disappears is the 
difference between the mass of the parent meson and the combined 
mass of the particles arising from its decay. The energy released 
appears as the kinetic energy of these particles. 

Considering the decay of * mesons first, the only visible track 
coming out of the end of a *■ meson track was that of a M meson. 
But there had to be at least one other particle born in the process. 
The principle involved here is conservation of momentum; the same 
principle is at work when a gun recoils in the opposite direction 
from that in which the shot is fired. The r meson, when it decays, 
is at rest and therefore has no momentum. The decay n meson, on 
the other hand, does have a momentum, which must be exactly 
canceled by an equal and opposite momentum carried by one or 
more "invisible" particles. This invisible particle carries no charge, 
or it would leave a track in the emulsion. 

Suppose that in addition to the y. meson only one invisible 
particle is born in a 7r-meson decay. Then there is only one way for 
this particle and the n meson to share the available kinetic energy 
so as to acquire equal and opposite momenta. But if two or more 
invisible particles arise from each x-meson decay, the decay 
products can share the available energy in many different ways 
without violating the conservation of momentum or the conserva- 
tion of energy (Fig. 9-6). Now according to the experimental data, 
whenever a decay M meson stopped in an emulsion, it always 
traveled the same distance from the point of origin (0.63 mm). 
Since the ft meson always acquired the same energy (4.17 MeV), 

Pi mesons 


the total energy was shared in only one way and therefore there 
was only one invisible particle. 

What was this neutral particle? The property most easily 
determined was its mass. According to Einstein's law, when the 
jr meson decays, it releases an amount of energy equal to nuc 1 



Fig. 9-6 (a) A charged particle that decays into another charged particle 
(heavy arrow) and a neutral particle (light arrow). The parent particle was at 
rest and therefore had zero momentum. Conservation of momentum requires 
that the momenta of the two "daughter" particles be equal and opposite, so 
that they add up (vcctorially) to zero. The kinetic energies of the two particles 
must add up to a fixed value equal to the velocity of light squared times the 
mass difference between the parent and the daughter particles. These two 
conditions determine in a unique manner the energies acquired by each of the 
two daughter particles. Thus, the charged daughter particle has the same 
energy in all decay events. The mass of the unknown neutral particle can be 
calculated from this energy. (6.) and (6.) A charged particle that decays into 
another charged particle and two neutral particles. The momentum of the 
charged daughter particle must be equal and opposite to the vector sum of 
the momenta of the two neutral particles. The kinetic energies of all three 
must add up to a fixed value equal, as before, to the energy released by the 
disappearance of mass in the decay process. These two conditions, however, 
do not uniquely determine the energies acquired by each of three daughter 
particles. Consequently, charged particles arising from various decay events 
will have a variety of energies. The energy will be small if the two neutral 
particles are emitted in nearly opposite directions (6i) and large if they are 
emitted in nearly the same direction (b,). The energy of the charged daughter 
particle is maximum when the two neutral particles are emitted in exactly the 
same direction. From this maximum energy it is possible to calculate the com- 
bined mass of the two neutral particles. 


Cosmic Rays 

(m, is the ir-meson mass). Of this energy an amount equal to m^c* 
appears as the mass m„ of the p meson. Another known amount 
(4.17 MeV) is accounted for by the kinetic energy of the y. meson. 
It is a simple matter to determine the amount of energy left over 
(about 30 MeV, to be specific). Because of conservation require- 
ments, this amount must be equal to the energy moC 5 needed to 
create the (unknown) mass m of the neutral particle and to give 
the particle whatever (unknown) kinetic energy it may have. We 
can now make calculations for different masses. For each assumed 
mass we can compute m^c 1 ; the energy left over is the kinetic 
energy. From the kinetic energy and mass we can compute the 
momentum. If we have made the right guess, the momentum of the 
neutral particle must equal the momentum of the n meson (which 
can be computed from the known values of its mass and its kinetic 

An analysis of the kind outlined above showed that the mass 
of the neutral particle is much smaller than the mass of the elec- 
tron, and most likely is zero. There are two neutral particles with 
zero mass, the photon and the neutrino. Because of their very weak 
interaction with matter,' neutrinos are exceedingly difficult to 
detect, but photons can be detected without too much trouble 
through the secondary electrons they generate in matter by pair 
production or the Compton effect. Failing to find photons among 
the decay products of x mesons, physicists concluded that the 
neutral, massless particle arising from the decay of a x meson was 
a neutrino. The decay of positive and negative x mesons can be 
described symbolically by the formulas: 

x+ -» M + + V 

*~-*lT + v 
where v represents a neutrino and v an antineutrino. 2 

In the case of \i mesons, many experimenters undertook to 
measure the energies of the electrons arising from the decay of 

1 See Appendix H. 
1 See Appendix I. 

Pi mesons 


"natural" n mesons in cosmic radiation and of "artificial" p mesons 
produced in high-energy accelerators. In 1949 Robert B. Leighton, 
Carl D. Anderson, and A. J. Seriff at the California Institute of 
Technology proved conclusively that the energies of the decay 
electrons were distributed over a wide range extending all the way 
from zero up to about 50 MeV. This meant there were at least Iwo 
invisible particles among the decay products. 

By that time physicists had gathered evidence that no photons 
were produced in the decay of m mesons. In addition, some of the 
decay electrons had such a large energy as to require a combined 
mass for the "invisible" particles of practically zero. (Figure 9-6 
clarifies this argument.) Therefore the neutral particles must 
again be neutrinos, and theoretical arguments showed that there 
were just two neutral particles. The decay schemes of positive and 
negative n mesons then can be described symbolically by the 

H+ — » e* + v + v 
p- — ♦ e~ + v + v 
where e represents an electron. 

The discovery of the neutral pi meson 

To complete the x-meson story, I should mention one other im- 
portant discovery, a discovery that resulted from the combined 
efforts of cosmic-ray physicists and physicists working with high- 
energy accelerators. 

Ever since the discovery of positive and negative mesons, 
physicists had speculated about the existence of neutral mesons. 
They had rather foggy ideas about what the properties of these 
particles might be until, in 1947, the problem was brought into 
sharper focus by Oppenheimer. He suggested that neutral mesons 
might decay very rapidly into photons and that the photons thus 
produced might be responsible for most of the cosmic-ray photons 
and electrons found in the atmosphere. 


Cosmic Rays 

Oppenheimer's suggestion proved correct. The first experi- 
mental indication of the existence of neutral mesons came from 
cosmic-ray observations. Cloud-chamber photographs obtained in 
1950 by the cosmic-ray group at the Massachusetts Institute of 
Technology provided strong evidence that high-energy nuclear 
interactions frequently gave rise not only to penetrating particles 
(charged *■ mesons) but also to neutral particles capable of ini- 
tiating cascades. These neutral particles could well be photons 
arising from the decay of neutral mesons. Observations with 
nuclear emulsions exposed to cosmic rays confirmed these results. 
The final and crucial evidence for the existence of neutral mesons, 
however, came from experiments with a high-energy accelerator 
performed in 1950 at the University of California in Berkeley. 

Subsequent experimental studies assigned a mass of 264 elec- 
tron masses to the neutral meson, a value very close to the mass of 
the positive or negative * meson. Hence, the neutral meson was a 
v meson rather than a p. meson. Supporting this view was the fact 
that neutral mesons, like charged w mesons, were abundantly pro- 
duced in high-energy nuclear interactions. The mean life of the 
neutral meson was estimated to be of the order of 2 X 10~ IS 
second. It was also found that the meson produces two photons 
when it decays. Its decay scheme, therefore, is: 

t° -* y + y 

where ir° represents a neutral v meson and y a photon (y ray). 1 
The mean life of t° mesons is about 100 million times shorter than 
the mean life of charged * mesons. Consequently, x° mesons decay 
so near the point at which they originate that the two photons 
arising from their decay usually appear to come exactly from that 
point. (The photon paths, though invisible, can be derived from 
the observed tracks of the electron pairs produced in the emulsion 
when the photons undergo materialization of energy.) Only t° 
mesons of exceedingly high energy live long enough (because of the 

1 There are other possible modes of decay for *° mesons, but they are 
exceedingly rare. 

Pi mesons 


relativistic effect discussed on page 117) to travel a distance of 
several microns before decaying (one micron is one-millionth of a 
meter). It was through accurate measurements of this distance in 
nuclear emulsions that it became possible to estimate the mean 
life of t° mesons. 1 

' Within the bounds of this book it is impossible to give proper credit to 
the many scientists who made important contributions to the discoveries 
related above. However, it is of historical interest to note that Japanese 
physicists, working in almost complete isolation during World War II, reached 
a number of conclusions that later proved to be correct. Even before the 
magnetic-lens experiment of Conversi. Pancini, and Piccioni (see page 127). 
Japanese physicists had become convinced that cosmic-ray mesons did not 
interact with atomic nuclei as strongly as those postulated by Yukawa. They 
argued that, if the interaction were strong, cosmic-ray mesons would not have 
been capable of traversing very large thicknesses of matter. As a way out or 
this difficulty, Tomonaga and his associates, in 1943, suggested that the cosmic- 
ray mesons and the Yukawa mesons might be different particles. (This 
two-meson hypothesis was formulated again by Robert E. Marshack and Hans 
A Bethe, who did not know of Tomonaga's work, shortly before the experi- 
mental discovery of x mesons.) Also, in 1942, H. Tamaki postulated the 
existence of neutral mesons and presented the hypothesis that they decayed 
into photons, thereby giving rise to the bulk of the soft component of cosmic 


Nuclear interactions 
of cosmic rays 


On several occasions I have referred to the nuclear interactions of 
cosmic-ray particles. The time has come to look into this matter 
more closely. What evidence is there to show that such interactions 
actually occur? What are they like? Which among the various 
kinds of particles found in the cosmic radiation are actually capable 
of producing them? The present understanding of these inter- 
actions, which is still incomplete, has come from more than a 
quarter century of experimental and theoretical investigation. 

Early evidence 

As I mentioned in Chap. 6, when showers were first discovered in 
the early 1930s, many of us thought they resulted from the colli- 
sions of cosmic-ray particles with atomic nuclei. In such collisions, 
supposedly, nuclei were broken up and many positive and negative 



Cosmic Rays 

electrons were produced simultaneously. Theories of such multiple- 
produclion processes were developed by a number of physicists, 
notably Werner Heisenberg in Germany and Gleb Wataghin in 
Italy. However, it soon became clear that showers required a very 
different explanation. 

Positive and negative electrons appeared in single pairs, not 
in groups of many pairs at a time. The large numbers of particles in 
the observed showers were the result of many separate events 
involving pair production by photons and radiation by electrons. 
These processes occurred in the vicinity of atomic nuclei, which 
provided the electric held necessary for pair production and 
radiation but were themselves left intact. 

Nonetheless, there were also reasons for believing that cosmic- 
ray particles did sometimes break up atomic nuclei and, in so doing, 
gave rise to groups of secondary particles entirely different from 
those of ordinary showers. 

The earliest evidence for such nuclear disintegrations came 
from the so-called cosmic-ray stars discovered by Marietta Blau 
and H. Wambacher of Austria in 1932. These stars, which appeared 
in photographic emulsions that had been lying about for a long 
time before development, consisted of groups of particle tracks 
diverging from a single point. Because they were working with 
emulsions in which only heavily ionizing particles would leave 
observable tracks, Blau and Wambacher attributed the tracks to 
comparatively slow protons and a particles with energies of the 
order of 10 MeV. They then attempted to determine whether the 
"stars" were produced by radioactive contamination of the plates. 
(A tiny speck of a radioactive substance, such as polonium, can 
produce a star of a particles appearing to come from a single point, 
even though the decay of each polonium nucleus gives rise to only 
one a particle.) However, Blau and Wambacher had to discard this 
interpretation, because the energy of the particles was too high, 
and they concluded that the stars were the result of nuclear dis- 
integrations caused by cosmic rays. Later, in confirmation of their 
conclusion, the stars were found to occur more frequently in photo- 

Nuclear interactions of cosmic rays 


graphic plates kept at high elevations, where the cosmic-ray 
intensity is greater. 

New experimental methods 

For several years cosmic-ray stars remained the only direct evidence 
for interactions between cosmic rays and atomic nuclei. The energy 
required to produce these stars was modest on the cosmic-ray scale 
(of the order of 100 MeV). Moreover, the stars were quite rare. 
Hence, it appeared that nuclear interactions did not play any 
significant role in the general picture of cosmic-ray phenomena. 
But the discovery of mesons changed the picture completely. 
Mesons had a short life, and therefore they could not come from 
any great distance. In other words, they were not part of the 
primary cosmic radiation but were born in the atmosphere. The 
question that then arose was how they were produced. 

Some physicists considered the possibility that photons might 
produce meson pairs by a materialization process similar to the 
production of electron pairs. However, it was generally believed to 
be more likely that the observed mesons were the result of nuclear 
interactions. (Pair production of mesons by photons does occur, as 
experiments performed many years later demonstrated, but the 
event is extremely rare.) The theoretical ideas about multiple pro- 
duction of particles in nuclear collisions, originally proposed to 
explain ordinary showers, were now revived and applied to mesons, 
and experimental physicists set out to search for these processes. 
Wataghin, then in Brazil, and Janossy, in England, were the 
first to begin this new line of research, in about 1940. One of the 
experimental arrangements used by Janossy and his collaborators 
is illustrated in Fig. 10-1. It consisted of a number of G-M counters 
with the appropriate electronic circuits to detect coincidences 
between counters out of line. In principle the method was very 
much like the one used for the detection of ordinary showers 
(Chap. 4). However, Janossy 's counters were so arranged that 
thick layers of lead could be placed between them. Although the 


Cosmic Rays 

presence of lead between the counters cut down the number of 
coincidences markedly, a significant number was still observed 
with as much as 50 cm of lead between the counters. These coin- 
cidences could not have been due to ordinary showers, since 
ordinary showers could not have traversed 50 cm of lead. They were 
therefore ascribed to groups of penetrating particles, presumably 
mesons, produced by nuclear interactions of cosmic rays in the 
material above the counters. 

Counter arrangements similar to the one shown in Fig. 10-1 
(known as penetrating-shower detectors) proved to be very conven- 
ient in the study of nuclear interactions of cosmic rays. Notice 

G 4 - 

Fig. 10-1 Penetrating-shower detector used by Janossy in his early experi- 
ments. Electronic circuits select simultaneous discharges of at least one counter 
in group G, and two counters each in groups G, to G 4 . Such coincidences can be 
produced only by showers containing penetrating particles capable of trav- 
ersing large thicknesses of lead. 

Nuclear interactions of cosmic rays 


that they will respond only to interactions initiated by particles 
of very high energies — many BeV at least — for it takes a large 
amount of energy to produce several particles capable of traversing 

50 cm of lead. 

At about the same time another very useful instrument, the 
multiplate cloud chamber to which I made incidental reference in 
Chaps. 7 and 9, was being developed. The multiplate cloud cham- 
ber, often rectangular in shape, contains a number of horizontal 
plates, usually metal. The purpose of the plates is to increase the 
probability that a cosmic-ray particle will interact with matter 
while traversing the chamber. In most cases, the chamber is trig- 
gered by an array of G-M counters so arranged as to favor the 
detection of such interactions. A multiplate chamber can be 
thought of as an absorber cut into slices; the particle tracks are 
visible in the space between slices and thus provide a fairly detailed 
picture of what actually happens when the particles traverse the 


In about 1939, J. C. Street (who had collaborated with E. C. 
Stevenson in making the first measurement of the M-meson mass; 
see Chap. 8) constructed a multiplate chamber for the purpose of 
studying ordinary photon-electron showers. Among large numbers 
of showers he observed a few events of a different type in which, 
apparently, cosmic-ray particles had collided with atomic nuclei. 
The distinguishing feature of these events was the production of 
secondary particles that passed through the metal plates without 
initiating electron showers; the particles were not electrons, then, 
but probably mesons. A typical shower containing mesons is shown 

in Fig. 10-2. 

A third instrumental development of very great importance in 
the detailed study of the nuclear interactions of cosmic rays was 
the revival in the middle 1940s of the nuclear-emulsion technique 
discussed in the preceding chapter. Nuclear emulsions sensitive to 
minimum-ionizing particles made it possible to detect the produc- 
tion of high-energy mesons in cosmic-ray stars. It was found, 
indeed, that when a particle of sufficiently high energy collides 

Nuclear interactions of cosmic rays 


with a nucleus, it often produces groups of many mesons, as some 
theorists had predicted (Fig. 10-3). 

It would have been difficult to establish this important fact by 
any technique that did not allow experimenters to "see" the nuclear 
interaction at its origin, as the emulsion technique did. In a multi- 
plate chamber, for example, the interactions occur within the plates 
and the particles arising from them must travel a certain distance 
before they appear in the space between plates. When many par- 
ticles emerge from a plate, one can never be completely certain that 
they were all produced in a single event and not in several suc- 
cessive collisions occuring next to one another. 

Nuclear-active particles 

The reader, at this point, may wonder which of the various kinds of 
particles found in cosmic radiation are responsible for nuclear inter- 
actions. Are all or only some of the particles capable of producing 
these interactions? Or are the interactions caused by particles of a 
still different kind that the experiments described thus far had 
failed to discover? 

These very questions began to worry physicists after they 
realized the importance of nuclear interactions in cosmic-ray 
phenomena. Some interesting clues came to light when they began 
to study the rate of occurrence of nuclear interactions under various 
absorbers and at different altitudes. For example, in 1947 D. H. 
Perkins reported that the number of cosmic-ray stars found in 
plates kept under a thick lead shield was not much smaller than 
that found in unshielded plates. Since the shield was sufficiently 

Fig. 10-2 Mesons in a penetrating shower traverse several metal plates of a 
cloud chamber without producing secondary showers. In plate 4, however, a 
secondary interaction gives rise to a high-energy photon or electron, which 
initiates an ordinary shower. The brass plates across the chamber are 1.25 cm 
thick. The photograph was made by the MIT cosmic-ray group in 1950. 


Cosmic Rays 

• ■ . • ■ : 

■ ■'■. . . .■■: *, ; *■ • 

■ •• . •■■. ■ \ . • 

/ •' '' --i 

J ■ Y \ yj; : . 

•;-: : ^- 

« 1 \ 

Nuclear interactions of cosmic rays 


thick to absorb practically all electrons and photons, some other 
particle or particles produced stars. 

In 1946 I had moved to the Massachusetts Institute of Tech- 
nology and there resumed my research on cosmic rays, which the 
war had interrupted. At one time or another in the following years, 
many distinguished scientists from the United States and other 
countries were associated with the MIT cosmic-ray group. Even a 
partial list would run to several dozen names. But it should be 
clearly understood that whenever I refer to the contributions of the 
MIT group, I am speaking primarily of work done by these scien- 
tists, and not of my own. 

Among the first problems to which the MIT group turned its 
attention was the nuclear interactions of cosmic rays. To determine 
whether p mesons were responsible for these interactions, John 
Tinlot in 1948 made a careful survey of the altitude variations in 
the rate of occurrence of nuclear interactions. Using a penetrating- 
shower detector (Fig. 10-4), Tinlot carried out measurements at 
sea level, at different locations in the Rocky Mountains of Colo- 
rado and at different elevations aboard a B-29 plane. His results 
showed that the number of high-energy particles responsible for the 
penetrating showers increased with altitude much more rapidly 
than the number of p mesons. For instance, between sea level and 
4 300 meters (the top of Mt. Evans), the rate of production of 
penetrating showers increased by a factor of 32 and the number of 
„ mesons by a factor of 2.5. Obviously, the particles that gave rise 
to penetrating showers in nuclear collisions could not be p mesons. 
Similar results were obtained at about the same time by other 
physicists using instruments designed to detect cosmic-ray stars. 

Kg 10-3 Star resulting from a nuclear collision of a high-energy particle 
(P) The star contains a number of tracks, among which those numbered from 
1 to 6 show minimum ionization. They belong to particles, presumably mesons, 
moving at nearly the speed of light. This is a typical example of mulUple- 
meson production. (From R. H. Brown, U. Camerim. P. H. Fowler W. 
Heitler, D. T. King, and C. F. PoweU, The Philosophical Magazine, vol. 40. 
p. 862, 1949.) 


Cosmic Rays 

What did these results mean? In the first place, they proved 
that high-energy electrons, photons, and M mesons, which together 
form the bulk of the cosmic radiation near sea level, do not interact 
appreciably with atomic nuclei. In the second place, they proved 
that the cosmic radiation in the atmosphere contains not only 




1,400 1,600 

400 800 1,000 1,200 

Atmospheric depth, g/crrT 2 
Fig. 10-4 Increase of penetrating showers with altitude as measured by 
Tinlot with the detector shown in the inset. Penetrating showers were detected 
by the simultaneous discharges of two or more G-M counters in each of the 
rows 6', to 6', (counters G, were not used in this particular measurement). 
The vertical scale gives the number of coincidences per hour; the horizontal 
scale gives the air mass above the instrument, in grams per square centimeter. 

Nuclear interactions of cosmic rays 


electrons, photons, and u mesons but also particles of one or more 
different kinds that are capable of breaking up nuclei and producing 
secondary mesons. These nuclear-active particles are very rare near 
sea level but they rapidly increase in number with altitude. 

The dimensions of atomic nuclei were known fairly accurately. 
From the cross-sectional area of nuclei and from the known number 
of atoms per gram, one could easily compute the probability that a 
particle traveling in a straight line would hit a nucleus while trav- 
ersing a given layer of matter. For a particle traveling a distance 
of 1 cm in iron, the probability was about 1 in 20. In the case of 
charged particles this estimate applied only if the particles had 
sufficiently high energy not to be deflected appreciably by the 
electric fields of nuclei. 

Cosmic-ray particles certainly satisfied this condition. And yet 
thousands upon thousands of cosmic-ray electrons, photons, and 
u mesons traversed 1 cm of iron without producing a single nuclear 
interaction. This meant that when one of these particles passed 
through a nucleus, nothing happened, either to the particle itself 
or to the nucleus. In other words, with respect to electrons, photons, 
and u mesons, a nucleus seemed to behave like a cloud penetrated 
by a bullet. The inability of u mesons to interact with nuclei was 
consistent with the fact, mentioned in the last chapter, that the 
mesons were not easily absorbed by nuclei when stopped in matter. 
The absence of nuclear interaction also explained why u mesons 
were absorbed so slowly in the atmosphere. Nuclear-active par- 
ticles, on the other hand, were absorbed very rapidly in the 
atmosphere (Fig. 10-4). In fact, the observed atmospheric absorp- 
tion was more or less consistent with the view that a nuclear-active 
particle disappeared whenever it happened to hit a nucleus of 

nitrogen or oxygen. 

The next problem was to determine the nature of nuclear- 
active particles. Various clues pointed to protons, to neutrons, and 
to t mesons as likely candidates. The cohesion of protons and 
neutrons in a nucleus, as I noted earlier, indicated that at close 
distances these particles exerted tremendous forces of attraction 


Cosmic Rays 

upon one another. Therefore, a proton or neutron should interact 
strongly with any nucleus it met. Pi mesons also should interact 
strongly with nuclei, since they were supposed to be produced in 
nuclear collisions. A number of experiments, which I cannot relate 
here in detail, established the correctness of these suppositions. 
The strongly interacting cosmic-ray particles found in the atmos- 
phere were indeed protons, neutrons, and those r mesons that had 
not already decayed into /i mesons. 

Hosts of new particles 

The rapid increase in the number of protons, neutrons, and r 
mesons with altitude made it very desirable for physicists interested 
in the nuclear interactions of these particles to carry out experi- 
ments at the highest possible elevations. In answer to this need, the 
sounding balloon technique was perfected to the point where it 
became possible to fly fairly heavy and elaborate scientific instru- 
ments for hours or days at altitudes above 100,000 feet. 

To accommodate experiments requiring ground-based instru- 
mentation, mountain laboratories sprang up throughout the world: 
the Laboratory of Pic du Midi at 2,860 meters in the French 
Pyrenees (near a well-known solar observatory of the same name); 
the Laboratory of Testa Grigia, near Cervinia, at 3,500 meters in 
the Italian Alps; the Inter-university High-altitude Laboratory of 
Echo Lake, near Denver, Colorado, at 3,240 meters in the Rocky 
Mountains; the Laboratory of Chacaltaya, near La Paz, at 5,200 
meters in the Bolivian Andes; and many others. Cosmic-ray experi- 
ments performed with balloons, in the mountain laboratories, and 
at sea level provided a wealth of information on high-energy 
nuclear interactions. Among the groups that distinguished them- 
selves in this work were those at the University of Chicago under 
Marcel Schein, the Ecole Poly technique of Paris under Louis 
Leprince-Ringuet, and the University of Bristol under C. F. 

Among the effects of high-energy interactions discovered by 

Nuclear interactions oj cosmic rays 


Fig. 10-5 Tracing of a cloud-chamber 
picture obtained by Rochester and 
Butler in 1947, which they interpreted 
as showing the decay, in flight, of a 
neutral particle (represented by the 
dashed line) into two charged particles. 
(From a paper in Nature, vol. 160, 
p. 855, 1947.) 

cosmic-ray physicists, the most baffling was the production of an 
amazing variety of previously unknown, short-lived "elementary- 
particles. Where the * meson had filled a theoretical gap, the new 
particles created a totally unexpected and even staggering over- 
flow. Indeed, the problem of finding a place for such particles, as 
well as for the M meson, in the general framework of physics became 
one of the most pressing tasks of theoretical physics. 

In 1947, just a few months after the paper announcing the 
discovery of the r meson appeared, George D. Rochester and C. C. 
Butler of the cosmic-ray group at the University of Manchester 
published two remarkable cloud-chamber photographs. One of 
them showed the tracks of two charged particles diverging down- 

Fig. 10-6 Tracing of a cloud-chamber 
picture obtained by Rochester and Butler 
in 1947, which they interpreted as show- 
ing the decay, in flight, of a charged 
particle into one charged and one neutral 
secondary particle. (From a paper in 
Nature, vol. 160, p. 855, 1947.) 


Cosmic Rays 

ward from a point in the gas of the cloud chamber (Fig. 10-5). 
After the most careful consideration of all possible interpretations, 
Rochester and Butler were forced to conclude that both particles 
had originated in the decay of a neutral particle that had come from 
above and, of course, had left no visible track. 

In the second photograph a charged particle had, it seemed, 
suddenly changed direction while traversing the chamber (Fig. 
10-6). Rochester and Butler were unable to explain this picture 
except by assuming that the charged particle had decayed in flight 
and had given rise to a secondary charged particle and to an 
invisible neutral particle. More importantly, neither the neutral 
particle invoked to explain the Grst event nor the charged particle 
in the second could possibly be identifled as any known particle. 

Powell's group at Bristol reported another unusual event in 
1949. Working with nuclear emulsions, they found a particle with 
an apparent mass intermediate between that of the w meson and 
that of the proton. After coming to rest in the emulsion, the 

Fig. 10-7 Decay of a heavy meson into three x mesons, observed by Powell's 
group at the University of Bristol in 1949. The heavy meson K comes to rest 
at A. There it decays into two fast » mesons (tracks a and 6) and a slow nega- 
tive t meson, which comes to rest at B. The negative x meson is captured by a 
nucleus, which explodes, ejecting two heavily ionizing fragments (e and d) 
and, presumably, one or more neutrons, which leave no visible track. (From 
R. Brown, U. Camerini, P. H. Fowler, H. Muirhead, C. F. Powell, and D. M. 
Ritson, Nature, vol. 163, p. 82, 1949.) 

Nuclear interactions of cosmic rays 


particle apoeared to decay into three mesons. One of the mesons 
stopped in" the emulsion and left the signature characteristic of 
negative * mesons, a nuclear star (Fig. 10-7). 

It was difficult to draw any clear conclusions from these few 
isolated observations. But the evidence quickly began to multiply. 
Anderson and his collaborators at the California Institute of Tech- 
nology (working at sea level and at 3,200 meters altitude) obtained 
a substantial number of pictures confirming the findings of 
Rochester and Butler. In the following years cosmic-ray physicists 
throughout the world detected more and more "new" unstable 
particles in cloud-chamber photographs and in nuclear emulsions. 
For a while there was a great deal of confusion about the 
number and properties of the particles required to explain all the 
experimental data. Finally it became clear that there were two 
groups of new particles, to which the Cosmic-Ray Conference held 
at Bagnere de Bigorre, France, in 1953, gave the names heavy 
mesons and hyperons. Heavy mesons and hyperons may be neutral 
or electrically charged. Heavy mesons are lighter than protons but 
heavier than * mesons. Hyperons are heavier than protons. Heavy 
mesons and hyperons never occur singly. Pairs of heavy mesons 
may be produced in nuclear collisions by a direct process of 
materialization of energy. A high-energy proton or neutron may 
turn into a hyperon when it strikes an atomic nucleus, giving rise 
at the same time to a heavy meson. In this case part of its energy 
appears in the mass of the heavy meson and in the greater mass of 
the hyperon. 

The first of the two particles reported in 1947 by Rochester 
and Butler had been a neutral lambda particle (A) - a hyperon — 
which decayed into a proton and a negative » meson: 

A ->p + + r" 

The second particle found by Rochester and Butler had been 
positive sigma particle (2 + ) - another hyperon - which decayed 
into a neutron (n) and a positive r meson: 

2+ -* n + t+ 


Cosmic Pays 

The particle reported in 1949 by the Bristol group had been a 
heavy meson, positively charged (K+), which decayed according to 
the scheme 

K+ -»,r+ + x + + r~ 

One of the reasons it was so difficult to disentangle the experi- 
mental findings was that many of the new particles decay in several 
different ways; for example, the positive K meson can also decay 
according to any one of the following schemes: 


+ T° + *+ 
+ T+ 

>M + + V 

K + -> e + + *■» + » 

Some of these decay schemes were considered incompatible 
with one another, according to previously accepted theoretical 
ideas, and the difficulty was resolved only when one of these ideas 
was shown to be wrong. I am referring here to the discovery (due 
to the theoretical work by T. D. Lee and C. M. Yang and to the 
experimental work by C. S. Wu and her collaborators in 1956) that 
conservation of parity, unlike the conservation of energy or the con- 
servation of momentum, is not a law of nature. 1 

The completion of this very brief and sketchy account requires 
some mention of the antiproton and the anlineulron. The first is a 
negative particle having the same mass as the proton. It bears to 
the proton the same relation as that borne by the positive to the 
negative electron. As in the case of the electrons, mutual annihilation 
is the result of the encounter of a proton with its antiparticle. The 
antineutron, though neutral and having the same mass, is not 
identical with the ordinary neutron but is instead the counterpart 
of it as the antiproton is the counterpart of the proton. Thus, a 

'The law of conservation of parity involves rather subtle concepts. 
Crudely speaking, it postulates that a physical system and its mirror image 
(exemplified by a right-handed and a left-handed screw) must obey the same 
physical laws. 

Nuclear interactions of cosmic rays 


neutron and an antineutron, when they come together, also annihi- 
late each other. , 
Antiprotons and antineutrons had been predicted by Dirac s 
theory In 1954, cosmic-ray observations with cloud chambers and 
nuclear emulsions recorded two events that looked suspiciously 
like annihilations involving antiprotons. However, it was only 
through the work done with high-energy accelerators by hmilio 
Segre and Owen Chamberlain at Berkeley that, in 1956, the exist- 
ence of antiprotons and antineutrons became an experimental 

Neutrons in cosmic rays 

Nuclear collisions produce not only high-energy secondary particles 
but also substantial numbers of protons, neutrons, and a particles 
with energies of the order of millions of electron volts. Protons and 
a particles of such low energies are very rapidly absorbed through 
ionization losses. As a result, they never occur naturally in any 
great abundance. But neutrons, of course, do not lose energy by 
ionization. In successive collisions with nuclei, they gradually slow 
down and eventually become easy prey to nuclear absorption. The 
most voracious nucleus in the atmosphere happens to be that of 
nitrogen (A = 14; Z = 7). Consequently, most neutrons end their 
free lives by falling into nitrogen nuclei, each of which then emits a 
proton to become a nucleus of the radioactive isotope carbon 14 

(A = 14; Z = 6). 

The experimental study of slow neutrons in the atmosphere 
has proved to be of considerable interest in the general framework 
of cosmic-ray research. Among the physicists who made important 
contributions to this subject, I wish to mention here Serge Korff of 
New York University and John A. Simpson of the University of 
Chicago. Moreover, the production of carbon 14 by slow neutrons 
has had an important application in a field quite removed from the 

» See Appendix I for a list of the known "elementary" particles and a 
description of some of their basic properties. 



Cosmic Rays 

study of cosmic radiation: archeology. The age of a piece of wood, 
say, can be determined by measuring the ratio of carbon 14 to 
carbon 12 (the common, nonradioactive isotope) in a sample of the 
wood. All plants take in carbon dioxide in the process of photo- 
synthesis and "exhale" it in respiration. During the life of a plant 
or tree, then, the ratio of carbon 14 to carbon 12 will always be the 
same as that in the atmosphere. When the plant dies, however, 
the ratio decreases as carbon 14 decays. Since carbon 14 has a 
half-life of 5,600 years, the ratio will decrease by one-half in that 
time. Thus it becomes possible to make fairly reliable age deter- 
minations on a wooden object (or other once-living material) dating 
back twenty thousand years or more. This is the principle of carbon 
dating, which was developed in the 1940s by Willard F. Libby. 

What cosmic rays are 

and what they do 

in the atmosphere 


Between 1940 and 1950, as the nature and the properties or the 
various particles found in the local cosmic radiation were gradually 
puzzled out, a unified and self-consistent picture of all cosmic-ray 
phenomena began to emerge. And by the end of that decade 
physicists had found the answer to the original question raised by 
Hess's discovery in 1912: What is the primary cosmic radiation 
that rains upon the earth's atmosphere from outer space? 

The primary radiation 

In 1940, experimental studies of the effect of the earth's magnetic 
field on the incident radiation (Chap. 5) had already shown that 
most, and possibly all, primary cosmic rays are positively charged 



Cosmic Rays 

particles. A few years earlier some physicists had identified the 
penetrating particles observed near sea level with primary cosmic 
rays. By 1940, however, these penetrating particles were known to 
l>e mesons (that is, n mesons), which did not live more than a few 
microseconds. The mesons, therefore, could not be part of the 
radiation arriving from outer space, but had to be produced in 
the atmosphere by primary cosmic-ray particles of a different kind. 
Barring the existence of hitherto unknown, positively charged, 
stable particles, the primary radiation had to consist of positive 
electrons, or protons, or heavier nuclei, or a mixture of all three. 
But until 1940, experimental investigations provided no clear evi- 
dence on which to base a conclusion. Then in the early 1940s 
Marcel Schein and his group at the University of Chicago under- 
took a series of balloon experiments (up to altitudes of about 
70,000 feet, where the pressure is about one-thirtieth of an atmos- 
phere) the results of which convinced them that primary cosmic 
rays were not electrons. Judging from their experiments, the par- 
ticles found at very high altitude passed through several centi- 
meters of lead without producing showers as abundantly as high- 
energy electrons were known to produce. Moreover, they did not 
appear to be absorbed by lead as rapidly as electrons. On the basis 
of the evidence, Schein and his collaborators took the view that 
primary cosmic rays were probably protons. 

The experiments on the nuclear interactions of cosmic rays 
discussed in Chap. 10 greatly strengthened this conclusion; for the 
number of nuclear-active particles, which constituted a small frac- 
tion of the total radiation found near sea level, increased with 
increasing altitude much more rapidly than the number of elec- 
trons, photons, and n mesons. Indeed, it increased steadily up to 
I he highest elevations attained in the experiments, whereas the 
intensities of other components of the cosmic radiation reached a 
maximum and then decreased again (see Fig. 11-7). Through 
careful analysis and discussion of their experimental data, phys- 
icists became convinced that practically all primary cosmic rays 
were nuclear-active particles. 

What cosmic rays are and what they do in the atmosphere 165 

The nuclear-active particles found in the atmosphere were 
protons, neutrons, and *■ mesons. The only stable positive particle 
among these was the proton. It was thus natural to conclude that 
primary cosmic rays were protons. This conclusion proved to be 

correct — in part. 

Our story takes us now to the Universities of Minnesota and 
Rochester, where, soon after the war, strong cosmic-ray groups 
had been established by Edward P. Ney and by Hans L. Bradt and 
Bernard Peters, respectively. In 1948 the two groups, working in 
collaboration, sent a sounding balloon, carrying nuclear-emulsion 
plates, up to an altitude of 94,000 feet. When the plates were 
developed, they showed a number of very dense tracks, most of 
which passed through the entire pile of plates. From the grain 
density it was possible to estimate the rate of energy loss along the 
path and therefore the total energy spent by the particles in trav- 
ersing the plates. The conclusion was that the particles had energies 

of at least many BeV. 

At such large energies protons behave more or less as mini- 
mum-ionizing particles. Since the ionizing power (and hence the 
grain density) increases with the mass and with the electric charge 
of a particle of given energy, the observed particles either were 
heavier than protons or carried more than one elementary charge. 
Indeed, it soon became apparent that both the mass and electric 
charge were greater than the mass and charge or protons and that 
the particles were the nuclei of elements heavier than hydrogen 
stripped of their electrons. (The same balloon that carried the 
nuclear plates also carried a small cloud chamber. The cloud- 
chamber photographs, too, showed tracks of fast, multiply-charged 


The heavy nuclei found in the upper atmosphere, because of 
their high energy, were without question related to cosmic radia- 
tion. The physicists who discovered them presented several con- 
vincing arguments that they were not secondary particles, but 
belonged to the primary cosmic radiation itself. Perhaps the most 
direct line of reasoning rested on the observation of nuclei carrying 


Cosmic Rays 

a greater charge than that of the nuclei of atmospheric gases. 
Obviously these nuclei could not arise from the disintegration of 
gas nuclei in collisions with primary cosmic-ray particles. 

The emulsions used in the Minnesota-Rochester experiment 
were not sensitive enough to detect fast protons. When some time 
later more sensitive plates were flown, they showed, as expected, 
large numbers of proton tracks. In fact, it turned out that most of 
the primary radiation actually consists of protons. The relative 
abundance of the various nuclei is given in Table 11-1. 

While there was no longer any doubt that primary cosmic rays 
are for the most part protons and, to a lesser extent, bare nuclei of 
heavier elements, experimenters continued to search for high-energy 
electrons and photons in the primary radiation. Experiments 
performed at the Massachusetts Institute of Technology in 1948 
and at the University of Minnesota in 1952 gave negative results. 
In 1961, however, James Earl at the University of Minnesota and 
Peter Meyer and Rochus Vogt at the University of Chicago suc- 
ceeded in detecting primary electrons with an abundance of the 
order of a few electrons per hundred incident particles (which 
makes the electron flux intermediate between that of helium nuclei 
and the combined flux of all other heavier nuclei). In the same year 
William Kraushaar and George W. Clark at the Massachusetts 
Institute of Technology, using a detector carried beyond the atmos- 
phere by an artificial satellite, found some evidence (still uncon- 
firmed as this is written) for a (lux of primary photons with energy 
greater than about 50 MeV of the order of 1 or 2 per 1,000 incident 

The generation of secondary rays in the atmosphere 

Out of all this there arises a natural question about what physicists 
call the "genetic relation" between the various components of the 
local cosmic radiation. In other words: Through what chain of 
interactions does the primary radiation produce the many different 
kinds of rays observed at various levels in the atmosphere? 

What cosmic rays are and what they do in the atmosphere 167 

Table 11-1 Composition of primary cosmic radiation 

The table gives the number of different nuclei per 100,000 proton, in the 
primary radiation. Only primary particles with energies neater than 2.5 BeV 
L nacleon arc considered (i.e.. protons (A = 1) with greater than 
Z BeV; helium nuclei (A = 4) with more than 4 X 2 5 = 10 BeV; carbon 
nuclei (A = 12) with more than 12 X 2.5 = 30 BeV. and so on|. The dato arc 
taken from a paper by V. L. Ginzburg and S. I. Syrovatsky, Jo,*nal of Theo- 
retical Physics (Supplement), No. 20, p. 1, 1961. _ 

— ■ Relative 

y number 


Hydrogen (p) 
Helium (a) 
Light nuclei 
Medium nuclei 
Heavy nuclei 








I have already mentioned the extensive and remarkable obser- 
vations carried out by Erich Regener and his group at high 
altitudes and at great depths under water. It was one of Regener's 
collaborators, Georg Pfotzer, who in 1935 first flew an instrument 
designed to detect cosmic-ray particles traveling in a more or less 
vertical direction. Pfotzer's experiment was cleaner than those 
performed previously with electroscopes, because electroscopes 
record rays coming from all directions; and at a given altitude the 
effective thickness of the atmosphere is different for rays arriving 
at different angles to the vertical. 

Pfotzer's instrument was essentially a G-M telescope pointing 
vertically. Figure 11-1 shows the experimental arrangement. The 
curve in Fig. 11-2 gives the coincidence rate as a function of 
atmospheric depth.' The most striking feature of the curve, which 
closely resembles the shower curves discussed earlier (Fig. 7-1), 
was the maximum found at an atmospheric depth of about 90 
g/cm ! . It showed, more directly than any prior experiment had, 
that most charged particles found in the atmosphere are of second- 
ary origin. Indeed, the large increase in the counting rate from the 
top of the atmosphere to a depth of 90 g/cm* could only be ex- 
> See Appendix A. Atmospheric depth as a function of altitude is given 
in Fig. 11-8. 


Fig. 1 1 — 1 Gcigcr-Miillcr telescope used 
by Georg Pfotzer in 1933 to measure Ltiu 
intensity of cosmic rays traveling 
through the atmosphere in a vertical 
direction. A coincidence is recorded 
whenever a particle discharges one G-M 
counter in each of the three groups 
Gi, fi'j, and Cj. 

Cosmic Fays 





plained as the gradual buildup of secondary rays through the 
interactions of primary particles in air. 

For several years in the middle 1930s, my own group at the 
University of Padua in Italy was actively engaged in experiments 
designed to study the genetic relationship between the various 
secondary components. We used different arrangements of G-M 
counters and absorbers to sort out individual groups of particles. 
For example, a vertical telescope of G-M counters separated by 
thick blocks of lead selected penetrating particles (later identified 
as n mesons). A triangular array of counters under a lead plate 
detected high-energy electrons and photons that produced showers 
in the plate (Fig. 11-3). For the most part, our experiments 
consisted of comparing the counting rates of various detectors at 
different altitudes on mountains. Other investigators undertook 
similar experiments. Particularly valuable and illuminating results 
were obtained in the middle 1940s by a group of Italian physicists 
under Gilberto Bernardini. 

What cosmic rays are and whal Oiey do in the atmosphere 169 

It would be impossible for me to describe here in any detail 
the results of these experiments, just as it would be impossible to 
describe the subtle and often quite involved arguments used in 
their interpretation. I should like, however, to recall briefly how 
the general view of the problem gradually evolved. I shall do it by 
imagining the descriptions of cosmic radiation that one would have 
been likely to elicit, at different times, from a cosmic-ray physicist. 
Of course, not every physicist would have given the same answer, 
and I may not have been completely unbiased in my selection. 

1928 The primary cosmic radiation consists of high-energy 
photons; the charged ionizing particles ol>served in the 
atmosphere are secondary electrons arising from Compton 
collisions with these photons. {False) 
1931 The charged particles found in the atmosphere are, for the 
most part, primary cosmic rays. (False) However, among 
the observed particles there are also some secondary elec- 
trons produced in the atmosphere. (True) 


400 600 

Atmospheric depth, g/cm 



Fig 11-2 Vertical intensity of cosmic rays as a function of atmospheric 
depth, according to Pfolier. The horizontal scale gives the atmosphenc depth 
in grams per square centimeter; the vertical scale, the number of co.ncdences 
per minute recorded with the telescope shown in Fig. 11-1. 




Cosmic Rays 

The cosmic-ray particles found in the atmosphere include 
high-energy electrons and photons, which can be recognized 
by their property of producing showers. (True) They also 
include penetrating particles, whose nature is still a puzzle. 
The penetrating particles are probably part of the primary 
radiation. (False) Some of the electrons and photons are 
produced by collisions of the penetrating particles with 
atoms in the atmosphere. However, not all of them originate 
in this way, because experiments have shown that the 
number of electrons and photons increases with altitude 
more rapidly than the number of penetrating particles. (True) 
The penetrating particles are mesons and, since mesons are 
unstable, they must be produced in the atmosphere. (True) 
The decay of mesons gives rise to electrons; electrons 
initiate showers containing photons as well as electrons. 




Fig. 11-3 Experimental arrangements for selecting different components of 
local cosmic radiation, (a) Coincidences between three G-M counters, placed 
one above the other, measure the flux of both "soft" and "penetrating" 
particles (that is, electrons and n mesons, respectively). (6) Coincidences 
between three G-M counters, placed one above the other and separated by 
lead blocks with a total thickness of approximately 10 cm measure the flux of 
"penetrating particles" alone, (c) Coincidences between three G-M counters 
in a triangular array, with a lead plate about 1 cm thick above them, measure 
tho flux of shower-producing particles (high-energy electrons and photons). 

What cosmic rays are and what Ihey do in the atmosphere 


These electrons and photons increase with altitude more 
rapidly than mesons because the density or air decreases as 
the altitude increases, and the probability that a meson will 
decay in a layer of air of a given mass per unit area increases 
correspondingly. (True) Thus the altitude variation of 
electrons and photons is consistent with the assumption that 
these particles originate entirely from mesons, mainly 
through decay processes. (False) 
1945 The recent, more accurate data on the altitude variations 
of electrons and photons have once more changed our views 
concerning the origin or these particles. We now believe that 
they do not arise solely from the decay of mesons or from 
the collisions of mesons with atoms. Probably some of them 
are produced in the same nuclear interactions believed to 
be involved in the production of mesons. (True) 
1950 There are several types of mesons, and those constituting 
the penetrating component of the local radiation are of the 
mu variety. Mu mesons arise from the decay of charged 
r mesons. Although electrons and photons arise in part from 
the decay of M mesons, to a large extent they arise from the 
decay of neutral ■* mesons and from the subsequent produc- 
tion of showers by the y rays born in these processes. 
Charged and neutral r mesons are produced in the high- 
energy interactions of primary protons and heavier nuclei 
with nuclei of the atmospheric gases. (True) 

A summary 

All the pieces of the cosmic-ray puzzle have now fallen into place, 
and it may be useful to look at the picture they make (Fig. 11-4). 
Figures 11-6 and 11-7 summarize the most significant data on the 
primary radiation and the various components of the secondary 
radiation. In both figures the vertical scale measures the directional 
intensity of the various kinds of rays; Fig. 11-5 explains the exact 
meaning of this parameter. 

— k- 



-•._ r r- 



-V5 i r \. 

Mf i 


ST— « 





1 (U 


I— J 



o L_, 

Z — k- 


■9 a 

s I 


a ff 















-a a a 

CO -o -> 

a * 

■* V. 



H .9 

1 2. 

C - -. 

J 5 1 1 S 

f g SSI 

- a •! s 

-— . u rg H CO 

3-f f-S I 

§ B -S 






~ § 



8 J 



* 8 

9 2 

■2 2 3- 

3 B *• 9 


^ °- fi 2 





2 1 



a a e 

in a 



,gi = 

s a 

■O .Q 






= t- 2 .2. 

JV/ia* cosmic rays are and what they do in the atmosphere 



Fig. 11-5 Measurement of directional inten- 
sity. Consider a G-M telescope consisting of two 
identical counters of length / and diameter d. 
The distance D between them is large compared 
to I and to d. The number of particles per second 
passing through the upper counter is propor- 
tional to the cross-sectional area A = I X d. 
The fracUon of these particles that traverse the 
lower counter is also proportional to A and 
inversely proportional to the square of the 
distance D. Thus the coincidence rate is pro- 
portional to A*/D>. The quantity that, multi- 
plied by AVD', gives the coincidence rate is 
defined as the directional intensity per square 
centimeter per unit solid angle per second. For 
example, the directional intensity is 100 times 
the counting rate of two counters with the 
dimensions d = 1 cm and I = 1 cm and at a 
distance D = 10 cm. A = 1 cm X 1 cm - 
1 cm»; D> - 100 cm'; A'/D* = 1/100. 

Primary cosmic rays are protons and bare nuclei of heavier 
elements. Their energies are distributed over a broad spectrum 
extending to unbelievably large values (Fig. 11-6). As these par- 
ticles approach the earth, their trajectories are bent by the earth s 
magnetic field. Particles of all energies can enter the atmosphere at 
the magnetic poles. Moving from the poles to the equator, however, 
particles of increasingly high energy are prevented by the magnetic 
Seld from reaching the earth (latitude effect). At the same time, 
magnetic deflection creates an asymmetry in the distribution of the 
arrival directions of cosmic-ray particles entering the atmosphere. 
Being positively charged, these particles arrive with greater 
abundance from the west than from the east (east-west effect). 

Upon entering the atmosphere, cosmic-ray particles soon col- 
lide with the nuclei of the atmospheric gases. The exact point at 
which the first collision occurs is, of course, a matter of chance. 
On the average, protons collide after traversing about 70 g/cm , or 
about one-fourteenth of the total air mass alx,ve sea level; « par- 


Cosmic Hays 

E, eV 

Fig. 1 1 -6 Energy spectrum of primary cosmic rays. Plotted on the horizontal 
axis is the energy E of cosmic-ray particles, in electron volts. Plotted on the 
vertical axis is the directional intensity (see Fig. 11-5) for particles of energy 
greater than E striking the atmosphere above the geomagnetic poles (where the 
effect of the earth's magnetic Held is negligible). The portion of the curve 
between 10» and 10" eV was obtained from balloon measurements of the 
geomagnetic-latitude effect. Observations with nuclear emulsions exposed at 
balloon altitudes have provided some data in the energy region between 10" 
and 10" eV (the particle energy being estimated from the characteristics of 
their nuclear interactions). All data beyond 10" eV come from experiments on 
air showers (Chap. 12). Note that the energy varies from about 1 BeV (10» eV) 
to about 10 billion BeV (10- eV). Actually, the largest energy recorded, as of 
August, 1962. was 6 X 10" eV (equal te the energy required to lift a mass of 
about one kilogram to a height of one meter). Note also the enormous range 
or intensity, which varies from about 0.15 particle per square centimeter per 

™« „ o. 8 "* 10 **"■ SCCOnd at 10 ' eV to a value about ,0 " «mes smaller at 
10" eV. Since ft can be shown that the total flux of particles from all directions 
above the horizon is x times the directional intensity, we find that a detector 
of 2 cm s at the top of the atmosphere will record about one primary particle 
per second and that the whole atmosphere (about 10" cm' in area) will receive, 
every second, about two or three particles with energies above 10" eV. 

What cosmic rays are and what Ihey do in Ihe atmosphere 175 

tides collide after traversing, on the average, about 25 g/cm»; and 
heavier nuclei collide after traversing still smaller thicknesses. 
Consequently, the probability that a primary cosmic-ray particle 
will escape nuclear collision until it reaches sea level is practically 
nil. At the highest accessible altitudes on mountains one will find a 
few surviving primary protons. But primary a particles and heavier 
nuclei are found only near the top of the atmosphere and thus 
require balloon-borne experiments for their detection. 

When a primary cosmic-ray particle collides against a nucleus, 
the target nucleus usually breaks up, as does the cosmic-ray particle 
itself if it is an « particle or a heavier nucleus. Among the dis- 
integration fragments are protons and neutrons, which often ac- 
quire sufficient energy to produce violent effects when they, in 
turn, collide with other nuclei in the atmosphere. 

But that is not all. Much of the energy of the primary particle 

goes into the creation of new, short-lived particles such as * mesons, 

heavy mesons, and hyperons. If the energy is great enough, pairs of 

protons and antiprotons, neutrons and antineutrons are also 

produced. Among all of these secondary particles, r mesons are the 

most abundant (at least in the range of energies typical of most 

primary cosmic radiation). Pi mesons (charged and neutral) are 

just about as effective as protons and neutrons in producing nuclear 

interactions, but nearly all neutral mesons decay before they have a 

chance to produce such interactions. On the other hand, charged 

r mesons, having a much longer mean life, often collide before they 

can decay. Also, according to the theory of relativity, the mean 

life of a moving meson increases considerably as the speed of the 

meson approaches the speed of light. Therefore, the probability 

that a * meson will live long enough to collide with a nucleus 

increases with the energy of the meson. 

Secondary protons, secondary neutrons, and secondary r 
mesons, along with other less abundant secondary products and 
whatever primary particles may survive, account for the nuclear 
interactions observed at any given level in the atmosplierc. The 
short average distance between nuclear collisions, the large energy 
degradation that every successive collision brings about, and the 


Cosmic Rays 

fast disappearance of r mesons by spontaneous decay explain why 
the number of nuclear-active particles decreases with increasing 
atmospheric depth as fast as it does (Fig. 11-7). 

Neutral r mesons decay promptly into photons. Photons soon 
disappear in a materialization process that produces pairs of posi- 
tive and negative electrons. The electrons radiate more photons. 






"5> 0.I6 




§ 0.12 




o- 0.08 


















c dept( 

'/ 9/= 





Fig. 1 1-7 Vertical intensity of three components of local cosmic radiation as 
a function of atmospheric depth, in grams per square centimeter. Curve NA 
represents nuclear-active particles with energies greater than about 1 BeV. 
Curve E represents electrons with energies greater than about 100 MeV. 
Curve M represents M mesons with energies greater than about 200 MeV. 
Atmospheric depth as a function of altitude is given in Fig. 11-8. 

What cosmic rays are and what they do in the atmosphere 177 


10 IS M 

Altitude, km 

Fig. 11-8 Atmospheric depth, in grams per square centimeter, as a function 
of altitude above sea level, in kilometers. 


Cosmic Rays 

Thus through a cascade process, whose alternate links are pair 
production and radiation, the main body of the electron-photon 
component of cosmic rays develops. Starting from zero at the top 
of the atmosphere, the intensity of the electron-photon component 
reaches a maximum at an atmospheric depth of the order of 100 
g/cm 2 and then decreases rapidly (Fig. 11-7). This decrease occurs 
because the photons originating from neutral x mesons produced 
in the upper atmosphere dissipate their energy in showers of low- 
energy electrons and photons. At lower altitudes no fresh supply 
of high-energy photons is available, owing to the rapid rate of 
decrease of nuclear interactions with atmospheric depth. 

The n mesons arising from the decay of charged ir mesons do 
not interact with nuclei, but pass freely through them. Hence, they 
loose energy only by ionization, and they disappear only by decay. 
The ionization loss in the atmosphere is comparatively small, and 
the mean life of n mesons is comparatively long. Therefore, neither 
process is very effective in eliminating /x mesons before they reach 
sea level. This accounts for the fact that the number of n mesons, 
after reaching a maximum in the upper atmosphere, decreases 
slowly with increasing atmospheric depth (Fig. 11-7). As a conse- 
quence, ii mesons become the dominant component of the local 
cosmic radiation near sea level. In fact, n mesons bury themselves 
deep into the ground, and some have been observed under as much 
as 1,000 feet of rock. The n mesons that decay in the atmosphere 
contribute some of the electrons of the local cosmic radiation. 

Giant showers 
of the atmosphere 


The life history of a cosmic-ray particle and its progeny outlined 
at the end of the preceding chapter applies to primary particles of 
all energies. But, as Fig. 11-6 shows, the energy spectrum of 
primary cosmic rays covers an enormous range, and there is an 
important difference in quantity, if not quality, between the 
observable effects of low- and high-energy particles. 

The family of secondary particles initiated by an "average" 
primary particle (with an energy, say, of 10 BeV) dies out after a 
few generations. Nowhere in the atmosphere does this family 
include any large number of members. At sea level a single particle 
(ordinarily a y. meson) may represent the last of the •■line," or the 
family may have disappeared altogether. By contrast, a primary 
particle with an energy, say, a million times greater - 10 million 
BeV — gives rise to a family much better equipped to survive. 
Although the particles of each new generation have considerably 
less energy than those of the preceding one, the original "capital" 



Cosmic Rays 

is so great to begin with that it takes a large number of generations 
before the energy received by each new particle falls below the 
limit where further production of secondary particles becomes 
impossible. Thus the family will grow to an enormous size before 
it begins to die out, and at sea level its membership will still num- 
ber in the millions. All of these particles, coming from a single 
primary cosmic ray, form what is known as an air shower. 

The discovery of air showers 

Air showers were discovered, more or less by chance, through the 
widespread application of coincidence-counter arrangements to the 
experimental study of cosmic rays. The devices used to detect coin- 
cidences will record as simultaneous the pulses of two or more 
counters if these pulses arrive within a certain small time interval. 
This interval, the resolving lime, was of the order of 0.01 second in 
the early experiments of Bothe and Kohlhorster. The development 
of vacuum-tube circuits of increasing sophistication eventually 
reduced the resolving time to considerably less than 1 micro- 
second. But, however short the interval, there is always a possi- 
bility that unrelated particles will cross the counters in such quick 
succession as to produce a coincidence. 

After physicists began to experiment with coincidences, it 
became a common practice to test the operation of the equipment 
by placing the counters out of line, usually on a horizontal plane. 
Then there could be no true coincidences caused by a single par- 
ticle traversing all counters. And without any heavy material above 
the counters, the number of true coincidences resulting from showers 
produced locally was negligible. Several experimenters must have 
noticed that the number of coincidences recorded under these 
circumstances was too large to be accounted for entirely by chance. 
I know I did, and I also noticed that the unexplained coincidences 
were more abundant at high altitude than at sea level. From these 
observations I concluded, if I may be forgiven for quoting from 

I Ol 

Giant showers of the atmosphere 

one of my own papers, "It would seem that occasionally very 
extensive groups of particles arrive upon the equipment. I hat 
was in 1934. Gradually the idea began to emerge that these very 
extensive groups of particles" were the result of cascade processes 
n the earth's atmosphere, just as ordinary showers were the result 
of cascade processes occurring in lead or other dense mater als. 
In 1938 the French physicist Pierre Auger and Ins collaborators 
undertook a systematic study that established beyond any doubt 
1 ocTurrence of air showers and provided some prehmmary 
"formation about their properties. These early expenments were 
carried out with relatively simple equipment that consisted of a 
ew G-M counters and a coincidence circuit. In one of their expen- 
ments Auger and his group used the three-counter setup dlustrated 
m Fig. 12-1. As they moved counter C farther and farther away 
from counters A and B, the coincidence rate gradually decreased 
(Fie 12-2) At the maximum distance between counters [lo 
meters), the apparatus still recorded coincidences, which showed 
that the air showers covered rather large areas. 

At first it was believed that air showers were m.ttated at the 
top of the atmosphere by high-energy electrons or photons and 
that the only processes involved in their ^0^^ 
tion losses by electrons and pair product.on by photons. Before 

,2-1 Air-shower experiments by Auger and hi. collaborate fa .1938 
were made with the counter arrangement shown here. Counters G and 6, 
were Z«l Z above the other, with their axes 22 cm apart. The UurdCQU^ 
proved horizontally * various distances 4 ranging from 15 cent.meters 
to 75 meters. 


Cosmic Hays 

long, however, mesons had been added to the list of air-shower 
particles. In the meantime, it was becoming increasingly evident 
that primary cosmic rays were neither electrons nor photons but, 
for the most part at least, protons and heavier nuclei. Obviously, 
an air shower was more than a simple electron-photon cascade. It 
was clearly a complex phenomenon in which nuclear interactions 
and spontaneous decay processes also played an important role. 

Here, then, is the general picture of an air shower. A high- 
energy primary particle (proton or heavier nucleus), upon entering 
the atmosphere, initiates a chain of nuclear interactions. The par- 
ticles born in these interactions are mainly high-energy protons, 
neutrons, and charged *■ mesons, which go on to produce other 
nuclear interactions (Fig. 11-4). They form the narrow "bundle" 
known as the core of the shower. At sea level the core is no more 
than a few meters in diameter. 

In each nuclear interaction a certain fraction of the energy is 
spent in the production of neutral mesons, which immediately 

Gianl showers of the atmosphere 


Fig. 12-2 Air-shower data obtained by Auger with the counter arrangement 
shown in Fig. 12-1. The horizontal scale gives the horizontal distance d 
between counter G, and the pair of counters & and G,; the vertical scale, the 
number of coincidences per hour. (From a paper in I* Journal dc Physique el 
Le Hadium, vol. 10, p. 39, 1939.) 

decay into photons. These high-energy photons, which occur 
throughout the shower core, initiate ordinary cascade showers and 
thereby give rise to large numbers of positive and negative elec- 
trons. Also, some of the charged * mesons decay into /. mesons 
instead of colliding with nuclei. Thus, p mesons also originate from 
the shower core. . 

Multiple scattering in the atmosphere causes shower particles 
to spread out laterally from the core. Such scattering represents the. 
cumulative effect of the many small-angle deflections that occur 
when charged particles pass through the electric fields of atomic 
nuclei. By the time the shower has traveled some distance through 
the atmosphere, its particles are distributed over a wide area. Their 
density is greatest at the center -the point where the primary 
particle would have hit if there had been no collisions - and it 
decreases gradually with increasing distance from this point. 

Air-shower experiments 

Since the middle 1940s the study of air showers has held a promi- 
nent place in cosmic-ray research for one reason in particular. Air 
showers are produced by those rare primary cosmic-ray particles 
whose energy is enormously greater than that generated in the 
most powerful particle accelerators. The minimum energy neces- 
sary to produce a shower observable at sea level is about 10" eV. 
The energy of the particles responsible for the largest showers ever 
observed is about 6 X 10" eV. These are indeed extraordinarily 
large energies for individual subatomic particles; 6 X 10 19 eV is 
about 10 joules, the energy required to lift a mass of 1 kilogram to 
a height of about 1 meter. 

Physicists, of course, are keenly interested in the properties of 
particles with such extreme energies. They cannot expect to pro- 
duce them in the laboratory in the foreseeable future. Existing 
accelerators reach a maximum energy of 35 BeV (3.5 X 10 10 eV), 
and the most ambitious plans for future machines aim at energies 


Cosmic Hays 

of the order of 1,000 BeV (10 12 eV). Consequently, physicists must 
turn to cosmic rays as the sole source of supply. To the astro- 
physicist the very existence of these particles poses formidable and 
fascinating problems. Indeed, what extraordinary processes are 
capable of accelerating particles to such enormous energies? In the 
hope of finding clues to the solution, they would like to know 
whether the most energetic particles come from all directions or 
only from certain regions of the sky. They would like to know what 
their energy distribution is; for example, how the number of par- 
ticles with energies between 10" and 10 18 eV compares with the 
number of particles with energies between 10' 8 and 10' 9 eV. 

There is no way of studying the high-energy region of the 
cosmic-ray spectrum other than by observing air showers. If 
physicists were to fly a detector with an area of 1 square meter 
on a satellite circling the earth above the atmosphere, they would 
not achieve their purpose. The detector would record only a few 
particles of energy greater than 10 16 eV in a single year. In a period 
of about 1 million years, on the average, the detector would record 
only one particle with an energy greater than 10" eV. 

Fortunately, the occurrence of air showers reduces the problem 
to manageable proportions. If the number of particles in a shower 
is sufficiently large (that is, if the primary energy is sufficiently 
great), the density of particles hundreds of meters from the core is 
still large enough to be measured with appropriate detectors. Thus 
one can observe all large showers striking an extensive area by 
distributing over this area a comparatively small number of 
detectors. In this manner one can compensate for the very small 
number of high-energy particles by using a very large area of 

Over the past decade and a half, air-shower experiments have 
been carried out intensively at many institutions throughout the 
world. A partial list includes the Lebedev Institute of Physics in 
Moscow, U.S.S.R.; the Institute for Nuclear Studies in Tokyo, 
Japan; the Tata Institute of Fundamental Research in Bombay, 
India; the Atomic Energy Research Establishment in Harwell, 

1 ft£ 

Giant showers oj the atmosphere 

England; the University of La Paz, Bolivia; the University of 
Sydney, Australia; and, in this country, Cornell University and the 
Massachusetts Institute of Technology. 

The instrumentation for the detection and study of air 
showers has become more and more sophisticated, and ,t has now 
reached the point where it is possible to obtam a great deal of 
prec^ information about each shower that strikes the equipment. 
To take a specific example, I shall describe here the expenmenta. 
method developed by the cosmic-ray group at MI 1 . 

The basic detectors are disks of fluorescent plastic, each about 
1 meter square and 10 cm thick. A shower particle passing through 
disk generates a brief flash of light. The flash ,s picked I up b, £ 
photomultiplier, which converts it into an electric pulse, five pulse 
travels along a cable to a central station, where ,t is amplified and 
d into a cathode-ray oscilloscope. The effect of the pulse isto 
deflect the electron beam of the osculoscope vertically. (The beam 
itself appears as a point of light on the oscilloscope screen.) At the 
same time the beam is rapidly swept across the screen producing 
the familiar oscilloscope trace. Consequently, the deflection caused 
by the pulse is recorded as a spike in the horizontal oscilloscope 

The fluorescent disks used in a given air-shower experiment 
feed their pulses to separate oscilloscopes, which are mounted one 
next to the other. Every time a shower strikes, the beams of he 
oscilloscopes are swept simultaneously, and a camera records the 
traces on all screens in a single photograph (Fig. U-i). 

At the time when the MIT program began, there was available 
a fair amount of information on the lateral spread of the shower 
particles around the core. However, no one had yet measured the 
longitudinal spread; that is, no one had taken a "snapshot showing 
the distribution of the shower particles along the direction of 
propagation. Since particles of all successive generations travel at 
nearly the velocity of light along paths that diverge only slightly 
from the direction of the primary particles, the longitudinal spread 
was believed to be quite small. George Clark of MIT and Pietro 

Fig. 12-3 Experimental arrangement used by the MIT cosmic-ray group to 
study air showers. Fluorescent plastic disks (thin rectangles at top) emit 
Hashes of light when struck by charged particles. At the center of each disk is n 
photomultiplier tube that converts the light into an electrical pulse; the ampli- 
tude of the pulse is proportional to the brightness of the (lash. Pulses travel to 
cathode-ray oscilloscopes (circles) through transmission lines containing delay 
circuits, which equalize the lengths of the electrical paths. Horizontal sweeps 
of all oscilloscope screens (grids) are triggered at the same time whenever 
three or more pulses pass through the coincidence circuit simultaneously. 
The amplitudes of the "spikes" (that is, the heights of the vertical deflections 
in the oscilloscope traces) indicate the numbers of particles striking the cor- 
responding detectors. The positions of the spikes in the horizontal traces show 
the relative arrival times of the particles. 

Giant showers of the atmosphere 


Bassi, a visiting scientist from Italy, proved that it is. They found 
that all shower particles went through each of their detectors 
within intervals of the order of a few times 10"' second. Since the 
velocity of light is 3 X 10 8 meters per second, this meant that 
the longitudinal spread was of the order of a meter or so. 

In other words, at a given instant, all shower particles are 
crowded within a flat disk about one meter thick. One can, then, 
visualize the propagation of a shower by thinking of this disk as 
traveling through the atmosphere at nearly the velocity of light 
while shower particles are continuously fed into it from the core 
and spread gradually outward. 

The results of Clark and Bassi's work suggested a convenient 
method for determining the direction of propagation of showers. 
Indeed, it is clear that, if a shower travels vertically downward, it 
will strike all detectors at the same time. If the shower comes from 
an inclined direction, then it will strike some of the detectors a 
little sooner, others a little later. Since the horizontal sweeps of the 
oscilloscopes are triggered simultaneously, the relative positions 
or the spikes in the oscilloscope traces are a measure of the time 
intervals between the pulses from the corresponding detectors 
(Fig. 12-4). By knowing the time when the shower first hits each 
detector, one can then compute the direction in which it was 


Fig. 12-4 Shower disk approaching detectors (represented by circles on a 
horizontal plane). 


Cosmic Rays 

traveling. This is also the direction of the primary particle that 
produced the shower. 

Furthermore, since the shower particles arrive almost simul- 
taneously, their pulses combine into a single pulse whose size is a 
measure of the density of shower particles at the corresponding 
detector. This information is the basis for determining the center 
of the shower and the total number of shower particles. The method 
is one of trial and error. Inspection of the record provides a pre- 
liminary estimate of the position of the shower center, the point at 
which the concentration of particles is greatest. Next, the number 
of particles recorded by the various detectors is plotted against the 
distance of the detector from the presumed center. Since the den- 
sity of shower particles decreases in a regular fashion with increas- 
ing distance from the center, the points should lie on a smooth 
curve. If they do not, one tries a second guess, using the observed 
deviations to figure out in which direction to move in order to 
obtain a better fit, and so on. 

Once the center is located and the curve giving the density of 
particles as a function of distance from the center is found, one can 
calculate the total number of particles in the shower. Actually, 
this whole process of successive approximations is carried out auto- 
matically by an electronic computer, and the complete analysis of 
a shower takes only a fraction of a minute. 

The method I have outlined here has been applied in a number 
of experiments performed at different altitudes, at different lati- 
tudes, and with different areas of detection. Between 1954 and 
1957, for instance, an experiment was carried out near Cambridge, 
Massachusetts, at sea level, with eleven detectors arranged within 
a circle 460 meters in diameter (Fig. 12-5). The largest shower 
detected in the experiment contained about 2.5 billion particles. 
The primary particle that produced it had entered the atmosphere 
at an angle of 11° from the vertical. The primary energy was about 
5 X 10 18 eV. The core struck somewhat outside the outer ring of 
detectors, at the point marked with a cross in Fig. 12-5. The figure 
also shows the number of shower particles that traversed each 
detector in this particular event. Figure 12-6 gives the number of 

Giant showers of the atmosphere 



Sb 12-5 Detector array used by the MIT group in an experiment at sea 
level (see text). The core of the largest shower observed struck at the point 
marked with a cross. The numbers of shower particles recorded by the various 
detectors on this occasion are indicated. 










200 300 400 

Distance, meters 




Fig. 12-6 The number of shower particles that traversed each detecter shown 
in Fig. 12-5 is plotted (on a logarithmic scale) against the distance of the 
detector from the center of the shower (cross in Fig. 12-5). 


Cosmic Hays 

+ 90 

+ 60 

+ 30 

Mercotor projection of celestial sphere 
30 60 90 130 150 180 210 240 270 300 330 340 












' _ - 

ry-;- y . 

i L S 




• -, 



■ * 

* ". 

s . • 

. •« ' \ 




k • •_' 

ft ■' 

' w "~ 1 



'• JJ ' »■* 

•i , 





« . 

■ • ■/■ 

■■' ■ . 

- .. 





, • ■ ■ 

^■-e 7 

-» r-*r» 

* • - 


, i 










: ~:z 



-- iV=1.8x10 7 





i evenrs 

Degrees 30 60 90 120 ISO 180 210 240 270 300 330 360 

Hours 2 4 6 8 10 12 14 16 18 20 22 24 

Right ascension 

Fig. 12-7 Directional distribution of the 652 largest showers recorded by the 
array shown in Fig. 12-5. Dots on this Mcrcator projection of the sky represent 
arrival directions. 

shower particles passing through each detector as a function of the 
distance of the detector from the core. 

One of the main purposes of the experiment was to determine 
from which regions of the sky primary cosmic-ray particles of very 
high energies come. The dots on the maps in Fig. 12-7 represent the 
arrival directions of the particles responsible for the 652 largest 
showers observed. Absorption by the atmosphere favors the detec- 
tion of showers coming from directions near the vertical. This 
explains why the dots in Fig. 12-7 cluster around the line at 42° 
declination, 42° being the latitude of Cambridge. Apart from this 
trivial effect, however, the data give no indication of any preferred 
direction of arrival for the primary particles. 

Another aim of the experiment was to determine the relative 
arrival rates of showers with different total numbers of particles. 
The experimental data, plotted in Fig. 12-8, show that the rate of 
arrival of showers decreases rapidly, but in a quite regular fashion, 
as the number of shower particles increases. There is no indication 

Gianl showers of Uie atmosphere 

of any maximum shower size. By further increasing the detection 
area or by continuing the observations over longer and longer 
periods of time, then, one should be able to detect showers with 
greater and greater numbers of particles. 

This surmise was tome out in an experiment carried out by 
John Linsley of MIT and Livio Scarsi, a visitor from Italy, between 
1959 and 1960 at an altitude of 1,700 meters near Albuquerque 
New Mexico. Nineteen detectors, each consisting of four fluorescent 
disks, were arranged in a heocagonal array 18 kilometers m diam- 
eter, as shown in Fig. IS* In a 6-month period abou 10,000 
showers of more than 10 million particles each were observed. 
Among them there were two of nearly 10 billion particles each. 
Linsley and Scarsi then expanded the array to form a hexagon 3.6 

io 6 


io» '0' 

Number of shower porticles 

Fi« 12-8 Frequency, in arbitrary units, of air showers as a function of the 
l!e!!f S ho:i P a7uc,e,Noti<»thevery rapid "-g^gX 
with increasing shower size (10' showers w,th about 1.7 X 10 
every 10 showers with about 2 X 10" particles). 


Cosmic Rays 

p ^j _q 

/ \ 

/ \ 

<i o o o 


b _o ^ 

1,800 or 3,600 meters 

Fig. 12-9 Detector array used by the MIT group in an experiment at 1,700 
meters altitude near Albuquerque, New Mexico (see text). 

kilometers in diameter. In the first few months of operation, the 
larger array recorded a shower containing ahout 30 billion particles, 
corresponding to a primary energy of ahout 6 X 10 19 eV. 1 


1 The data obtained from the experiments I have described here, as well 
as data from similar experiments, were used in plotting the energy spectrum 
of cosmic rays shown in Fig. 11-6. 

The Van Allen 
radiation belt 


On October 4, 1957, when Soviet scientists launched the first arti- 
ficial satellite into an orbit around the earth, the direct exploration 
of outer space became a concrete possibility. Cosmic-ray Phyaciste 
were quick to seize upon this opportunity. On November 3, 1957, 
the U S S.R. launched Sputnik II ; and the United States satellites 
Explorer I and Explorer III followed on February 1 and March 26, 
1958, respectively. All three satellites carried G-M counters for the 
purpose of making a preliminary survey of the cosmic-ray intensity 
at very high altitudes. The most striking result of these observa- 
tions was the discovery of a great radiation belt the 

earth. „ . 

Whether or not the particles in this belt are to be called cosmic 
rays is merely a matter of definition. In any case, the belt 
is such an important feature of the space environment of the earth 
and has such close connections with cosmic rays themselves that it 
deserves at least a brief mention in this book. 


!9* Cosmic Fays 

The discovery of the Van Allen belt 

The discovery of the radiation belt had its rools in scientific 
investigations dating from the early part of the twentieth century. 
In discussing the effects of the earth's magnetic field upon cosmic 
rays, I have already mentioned the theoretical work by Stormer 
on the trajectories of charged particles in the field of a magnetic 
dipole. Among other results, Stormer's computations had demon- 
strated that charged particles could be "trapped" indefinitely in 
this field (they are the particles traveling along the bounded 
trajectories mentioned in Chap. 5). This means that a particle 
injected with the proper momentum and in the proper direction 
will move forever within a limited volume of space surrounding the 
dipole (Fig. 13-1). The trajectories of the trapped particles are, in 
general, very complicated and very difficult to predict theoretically. 
However, they become comparatively simple when the momentum 
of the particles is sufficiently small. 

To understand the character of these trajectories, consider 
first the case of a uniform magnetic field. From Chap. 5 it will be 
recalled that the motion of a charged particle traveling at right 
angles to the magnetic lines of force descril>es a circle whose radius 


Fig. 13-1 Trajectory of a charged particle trapped in the magnetic field of a 
dipole, according to computations by Carl Stormer (1913). 

The Van Allen radiation belt 


(a) (M 

Fig 13-2 Motion of a particle in a uniform magnetic field B. (a) The particle 
moves along a circle in plane perpendicular to B while the plane moves at a 
constant speed V in the direction of B. (6) The trajectory is a spiral winding 
around a cylindrical tube of force. 

decreases as the momentum of the particle decreases or as the 
strength of the field increases (Fig. 5-2). If the particle is injected 
at any other angle to the field lines, it will spiral around a "tube or 
force" (Fig. 13-2). In simpler language, the particle rotates around 
a circle that lies in a plane perpendicular to the field lines while the 
center of the circle moves at a constant velocity along a field line. 
In a dipolar magnetic field such as the earth's, the lines of 
force are curved and the field strength varies from point to point 
(Fig. 5-1). Still, within a sufficiently small volume — a volume 
whose dimensions are small compared to its distance from the 


Cosmic Fays 

center of the dipole — the field will be approximately uniform. Now 
consider a particle with so small a momentum that the magnetic 
field makes it go around and around many times within the selected 
volume. Then the motion of the particle within this limited volume 
will closely resemble the motion in a uniform magnetic field. This 
means that, on a first approximation, the trajectories of charged 
particles are spiral paths around field lines. 

To go beyond these approximate conclusions, it is necessary to 
carry out a fairly sophisticated analysis of the problem. This was 
done by the Swedish astrophysicist Hannes Alfven in 1942. His 
computations led to the following description of the behavior of 
charged particles in a dipole magnetic field. Each charged particle 
travels in a circle the center of which {guiding center) oscillates 
back and forth in latitude between two symmetrically located 
mirror points along a curved guiding line. The guiding line has the 
shape of a line of force and rotates slowly about the axis of the 
dipole from east to west if the particle is positively charged and 
from west to east if it is negatively charged. As the guiding center 
oscillates along the guiding line — thus passing through regions of 
different magnetic field strength — the radius of the circle traced 
by the particle changes gradually in such a way that the circle 
encloses a constant magnetic flux (Fig. 13-3). 

Early in 1957 the American physicist Fred Singer published an 
article in which he called attention to the possibility that protons 
and electrons originating from the sun might occasionally become 
trapped in the earth's magnetic field. The opposite drift in longi- 
tude of the positive and negative trapped particles, Singer pointed 
out, would give rise to a sheath of electric current quite similar to 
the so-called "ring current" geophysicists had found it necessary 
to postulate in order to explain certain disturbances of the ter- 
restrial magnetic field (the main phase of magnetic storms; see 
Chap. 14). These, then, were the notions and the ideas current 
when Sputniks I and II initiated the exploration of outer space. 

The G-M counter aboard Sputnik II revealed nothing unusual. 
But the counters installed in Explorers I and Explorer III by the 

The Van Allen radiation bell 



Fig 13-3 Motion of a particle of small momentum in the earth's magnetic 
field (schematic). The particle moves in a circle; the center of the circle osc.l- 
lates back and forth along a guiding line, coincident with a line of force, be- 
tween two mirror points M and M'. At the same time the guiding line slowly 
rotates in azimuth (that is, about the geomagnetic axis). The magnetic flux 
through the circle (defined as the product of the area times the magnetic field 
strength at right angles to the area) is constant. Since the field strength vanes 
from point to point, this requires the area of the circle to change in accordance 
with the variations in the field. 

cosmic-ray group under James A. Van Allen of the University of 
Iowa behaved in a very peculiar fashion. At every revolution, the 
trajectories of Explorers I and III swung from a minimum altitude 
of several hundred kilometers (356 and 187, respectively) to a 
maximum altitude of several thousand kilometers (2,546 and 2,796, 

respectively). # 

Up to altitudes of about 1,000 km the counting rates showed 
only a modest and fairly regular increase. Above 2,000 km, how- 
ever, the counters apparently stopped working. Since the counters 
again transmitted pulses regularly when the satellites returned to 
lower altitudes, and since both satellites gave roughly the same 
results, it was hardly possible t« suspect a malfunction in the 
equipment. It was equally impossible to suppose that there were 
no cosmic rays al>ove 2,000 km. 


Cosmic Rays 

The only alternative explanation possible rests on the inherent 
"inertia" of G-M counters: if they are exposed to a radiation of 
excessive strength, they become "jammed" and stop counting 
altogether. Van Allen came to the conclusion that this was indeed 
the correct explanation of his results. According to his estimates, 
the intensity of the radiation responsible for jamming the counters 
must have been at least 15,000 times greater than the intensity of 
ordinary cosmic radiation. 

Van Allen announced his discovery at a meeting held in 
Washington, D.C., on May 1, 1958. In his report he stressed the 
fact that the radiation, though detected by G-M counters with 
fairly thick walls, could not penetrate within 600 km of the earth's 
surface. Since the total thickness of the atmosphere above 600 km 
is equivalent to only a minute fraction of the thickness of the 
counter walls, Van Allen concluded that the radiation must consist 
of charged particles prevented from approaching the earth's surface 
by the terrestrial magnetic field. He suggested that these particles 
originate from the sun, that they penetrate the trapping region of 
the earth's magnetic field by virtue of some sort of perturbation in 
the field, and that subsequently they remain trapped in this region. 

Sputnik II had failed to discover the radiation belt because its 
trajectory did not take it into the high-intensity zone while it was 
in radio contact with a receiving station. Immediately after Van 
Allen's announcement, the U.S.S.R. launched Sputnik III, 
equipped with various radiation detectors designed to record large 
radiation fluxes without jamming. These detectors counted at very 
high rates when the satellite passed through certain regions of 
space, thus confirming the existence of the high-intensity radiation 
discovered by Van Allen. 

In August of the same year, the United States launched 
Explorer IV, equipped with a variety of instruments specifically 
designed by the Iowa group to study the new radiation. At some 
locations these instruments revealed fluxes as high as 100,000 par- 
ticles per square centimeter per second. In agreement with the 
previous results, the intensity proved to be strongly dependent on 

The Van Allen radiation belt 


latitude. More important, the intensity varied with latitude 
measured from the geomagnetic equator rather than with latitude 
measured from the geographic equator. These observations con- 
firmed the view that the intensity distribution of the radiation was 
governed by the earth's magnetic field. 

Until that lime, no observations had been carried out at alti- 
tudes of more than a few thousand kilometers. Although the 
intensity had been found to increase steadily with increasing alti- 
tude, it was not known whether or not at still higher altitudes the 
intensity reached a maximum and then decreased. In other words, it 
remained uncertain whether the earth was actually surrounded by 
a radiation belt or whether it was immersed in a region of high- 
intensity radiation that might extend throughout the solar system. 
The deep space probes launched by the U.S.S.R. and by the United 
States in late 1958 and early 1959 provided the answer to this 
important question. They showed that the high-intensity radiation 
was confined to a belt around the earth and that at large distances 
(about fifteen earth radU or greater) only the normal cosmic-ray 

flux remained. 

These results (first announced by Van Allen — for whom the 
belt is named — and his collaborators and then verified by the 
Soviet scientists) left practically no doubt that the high-intensity 
radiation consists of particles trapped by the earth's magnetic 
field. Moreover, in the spring of 1959 a group of American scientists 
announced the results of an experiment, known as Project Argus, 
that confirmed in the most direct manner the theory of magnetic 
trapping. In the Argus experiment (suggested by the Greek-born 
scientist N. C. Christofilos prior to the discovery of the radiation 
belt and performed in the late summer of 1958) high-energy elec- 
trons were injected into the earth's magnetic field by means 
of small, rocket-borne atom bombs exploded at high altitude. 
The instruments aboard Explorer IV continued to detect these 
electrons, now trapped by the terrestrial magnetic field and 
spread into a thin shell around the earth, for a period of many 


The structure of the Van Allen belt 

Cosmic Bays 

Since 1958, a great deal of experimental data on the radiation 
belt has been obtained by means of the space vehicles already 
mentioned, by means of other space vehicles launched subse- 
quently, and also by means of sounding rockets shot into the lower 
edge of the radiation belt. Among those who made important 
contributions were American scientists working with Van Allen at 
the University of Iowa, with John Simpson at the University of 
Chicago, and with J. R. Winkler at the University of Minnesota 
and Soviet scientists, including V. I. Krassowsky, S. N. Vernov, 
A. E. Chudakov, and their associates. What follows is a summary of 
some of their most significant conclusions. 

The radiation bell appears to consist of electrons and protons. 
The energy spectrum is very steep, which means that the counting 
rate of an instrument sensitive only to particles with an energy 
greater than E decreases very rapidly as E increases. In fact the 
spectrum of the radiation belt is much steeper than the spectrum of 
ordinary cosmic rays, or, to put it another way, the average energy 
of the trapped particles is much smaller than that of cosmic rays. 
Furthermore, the spectrum changes from one region of the radia- 
tion belt to another. Consequently, the intensity distribution 
measured by a given detector depends to a large extent on the 
manner in which the detector responds to electrons and protons of 
different energies. 

A typical G-M counter detects, with nearly perfect efficiency, 
electrons of several MeV or more and protons of 30 to 40 MeV or 
more because these particles are capable of traversing the counter 
walls. It also detects, but with much lower efficiency, electrons with 
energies ranging down to some 20 keV. These electrons, upon 
hitting the counter walls, produce X-rays, some of which discharge 
the counter by producing Compton electrons. 

A G-M counter carried farther and farther from the earth will 
record two separate maxima in the radiation intensity (Fig. 13-4). 
Thus there are two distinct zones of high intensity separated 

The Van Allen radiation belt 

by a region of lower intensity. The intensity, as I have already 
noted, depends on the magnetic latitude as well as on altitude. 
Figure 13-5 shows a preliminary intensity map, drawn by \an 
Allen in 1959, in which there appears an inner bell in the neighbor- 
hood of the equator and an outer bell that reaches down toward the 

^The average energy of the particles is much higher in the inner 
belt than in the outor belt. The energies of protons in the inner 
belt range up to about 100 MeV; protons in the outor belt have 
energies of a few MeV at most. Moreover, the inner belt is com- 
paratively stable, whereas the population of the outer belt appears 
to fluctuate wildly, especially during periods of high solar activity. 

The origin of the Van Allen belt 

There is as yet no definite answer to the question concerning the 
origin of the radiation belt (or, rather, belts). However, it seems 





"g 1,000 




>>• — 


•> 100 



J ,0 

- J 



- j 




- j 








Fig. 13-* Counting rate of a G-M counter as a function of distance from the 
center of the earth, according to Van Allen (dotted hne corresponds to he 
1th'. surface). The data were obtained with a G-M counter earned by the 
Pioneer III space probe. 


Cosmic Rays 

likely that the proton component of the inner belt is largely a 
secondary effect of cosmic rays. Among the secondary particles 
arising from the collisions of primary cosmic rays with atomic 
nuclei in the atmosphere are neutrons. Some of the neutrons are 
ejected in an "upward" direction. They travel away from the earth 
undisturbed by the magnetic field because they carry no electric 
charge. But neutrons are unstable. They decay into protons and 
electrons after a mean life of about 16.6 minutes. Some of the 
neutrons will decay on their way out, thus injecting electrons and 
protons into trapped orbits. 


Fig. 13-5 Intensity map of the radiation belts, as measured with a G-M 
counter, according to Van Allen. The curves represent surfaces of constant 
intensity; for example, the counter ticks at a rate of 1,000 counts per second 
at all points of the surface of revolution generated by the rotation of the 1,000 
line around the geomagnetic axis. (Cosmic radiation is responsible for a rate 
of about 2 counts per second.) The dashed areas indicate the regions where the 
counting rate is greater than 10,000. The scale shows the distance from the 
center of the earth, in earth radii. (This map is the result of early experiments, 
which were analyzed under the assumption that the earth's magnetic field was 
exactly that of a dipole.) 

The Van Allen radiation belt 


From what is known today about the "upward flux" of neu- 
trons and about the mean life of trapped particles, it appears that 
this mechanism may account for most of the protons found in the 
inner belt. It is difficult, however, to account for all of the trapped 
radiation, and particularly that in the outer belt, in a similar 
manner. The fact that the structure of this belt is strongly affected 
by the activity of the sun lends considerable support to the idea 
that, in one way or another, the sun is involved in the production 
of the belt radiation. However, it is not clear whether the particles 
trapped in the outer belt actually come from the sun or whether 
they are accelerated locally as a result of mangetic disturbances 
originating from the sun. I shall have more to say about these 
disturbances in the next chapter. 

Cosmic rays 
and the sun 


One of the most striking result* of the early work on cosmic rays 
was the discovery that the radiation intensity appeared to be 
remarkably constant in time. Whatever small changes occurred 
were usually related to comparatively trivial atmospheric effects 
such as changes in atmospheric pressure. 

These early observations were carried out at or near sea level. 
There because of atmospheric absorption, only primary cosmic-ray 
particles of rather high energy (several BeV or more) produce 
observable effects. Later, when systematic measurements were 
undertaken at altitudes and latitudes where primary cosmic-ray 
particles of lower energy could also be observed, it became apparent 
that the low-energy portion of the cosmic radiatio.i was not at all 
constant in time and that its intensity changes had to do primarily 
with events in the sun. Even more spectacular changes in the par- 
ticle flux appeared when satellites and space probes took counters 
into regions where the shielding effects of the earth's atmosphere 
and magnetic field play no part. 



Cosmic Rays 

At the same time, measurements near sea level with instru- 
ments of increased accuracy were bringing to light small but 
significant changes in the cosmic-ray intensity which again were 
related to the activity of the sun. Moreover, it was found that once 
in a very great while there occurred, even at sea level, large tempo- 
rary changes in the cosmic-ray intensity. These changes in the flux 
of charged particles reaching the earth from outer space are an 
important element in the broad picture of the sun-earth relation- 
ship, which also includes such diverse effects as magnetic storms, 
aurorae, and blackouts in radio communication. 

Solar-flare particles 

The most spectacular solar event ever recorded by ground-based 
instruments took place on February 23, 1956. On that day, a great 
flare suddenly appeared on the face of the solar disk. A few minutes 
later the counting rates of cosmic-ray detectors all over the earth 
began to increase rapidly. In 15 to 20 minutes the counting rates 
had reached a maximum and had begun to decrease ; in a few hours, 
everything was over and the counting rates had returned to near 
their normal value. 

Various detectors located at stations all over the world gave 
quite different results. The largest effects were recorded by instru- 
ments located at fairly high latitudes and designed to detect the 
secondary neutrons generated by cosmic radiation (Fig. 14-1). 
The maximum counting rate of some of these detectors had in- 
creased to between 25 and 50 times normal. At the other extreme, 
/t-meson detectors located near the equator had recorded increases 
of only a few per cent. 

There is no doubt that the observed increases were due to a 
stream of high-energy particles, presumably protons, ejected by the 
sun at the time of the solar flare. Since even detectors located at 
low latitudes had recorded an increase, some of these particles 
must have had sufficiently high energies to penetrate the earth's 

Cosmic rays and me sun 


magnetic field near the equator, energies greater than 10 to 20 BcV. 
On the other hand, the effect was much smaller near the equator 
than at high latitudes, where particles down to one or a few BeV 
could penetrate the field. This meant the radiation originating 
from the flare contained comparatively fewer high-energy particles 
and comparatively more low-energy particles than did the ordinary 
cosmic radiation. The fact that the neutron detectors had recorded, 
in general, a much greater increase than had /(-meson detectors 
pointed to the same conclusion, because ft mesons are produced 
abundantly only by protons with energies of many BeV, whereas 
protons of lower energies are quite effective in producing neutrons. 
Four solar-flare increases, smaller in magnitude but roughly 
similar in character to the great increase of 1956, had been observed 



0500 0600 

Hours, universol time 




Fig M-l Increase in the counting rate of a neutron detector at Chicago, 
following the solar flare of February 26, 1956. (From John Simpson. Proceed,,** 
of (he National Academy of Sciences of the UniUd Stales of America, vol. 43, 
p. 42, 1957.) 

Cosmic Rays 

previously, in the years between 1942 and 1956. Several additional 
events, some fairly large, some quite small, were observed in the 
following years, at a rate of about one a year. In the meantime, 
however, it became apparent that while very few solar flares are 
capable of accelerating protons to energies of several BeV, the 
production of protons with energies of 100 MeV or less is a much 
more common event. 

Because of the earth's magnetic field, protons of this relatively 
low energy can approach the earth only near the poles. Even there 
they do not reach ground-based detectors, because energy losses 
through collision bring them to a stop in the upper layers of the 
atmosphere. However, their arrival increases the normal ionization 
of the upper atmosphere to a considerable extent, making it opaque 
to radio noises coming from the galaxy. The increased ionization 
also disrupts short-wave radio communications in the polar regions. 
It was the observation of these effects that provided the first 
evidence for the frequent arrival in the atmosphere of solar 
protons of relatively low energy. 

In 1956 and 1957, as part of the program for the International 
Geophysical Year, scientists in several countries set up a number of 
sensitive radio receivers to monitor the level of cosmic radio noises, 
and thereby to measure the radio opacity of the upper atmosphere. 
These instruments, called relative ionospheric opacity meters, or 
riomelers, frequently recorded a sudden increase in atmospheric 
ionization over the polar caps an hour or so after the appearance 
of a large solar flare. Because the increase was limited to the polar 
regions, these scientists attributed it to charged particles origi- 
nating, presumably, from the solar flare. The travel time from the 
sun to the earth and the depth of penetration into the atmosphere 
indicated protons with energies up to at least 10 MeV. 

The first direct check of these conclusions came in 1958. On 
August 21 of that year, Kinsey Anderson of the University of Iowa 
launched a balloon equipped with radiation detectors from Fort 
Churchill, Canada. The balloon reached a maximum altitude of 32 
kilometers at 10:00 p.m. the same day, and it remained at that 

Cosmic rays and the sun 


altitude until 5:00 p.m. the next day. At first the detectors counted 
at the normal rate. At 9:30 a.m. of August 22, however, the count- 
ing rate began to increase irregularly to a level about 10 times 
greater than normal. A few hours later the counting rate began to 
decrease slowly. As Anderson later learned, a large solar flare had 
occurred about 75 minutes before the balloon-borne detectors 
showed an increase in counting rate. From the thickness of the 
atmospheric layer above the balloon, he calculated that the protons 
recorded must have had energies greater than 100 Me\ . 

While these balloon observations were being made, the satellite 
Explorer IV was still orbiting the earth and making measurements. 
At 300-km altitude in the polar regions, the instruments aboard the 
satellite recorded the arrival of even stronger fluxes of solar par- 
ticles The effect was greater because the satellite's orbit kept it 
outside the earth's atmosphere, where the instruments could there- 
fore detect particles of lower energy than those capable of reaching 
Anderson's balloon. On many subsequent occasions, satellite-borne 
counters detected the arrival of solar particles after the appearance 
of solar flares. There was a close correlation between these observa- 
tions and those made with riometers on the ground. It thus became 
clear that one could confidently take the increased absorption ol 
cosmic noise, detected by riometers in the polar regions, as an 
announcement of the arrival of solar particles. 

The observation of solar particles provides very valuable infor- 
mation concerning both the solar atmosphere, where the particles 
supposedly originate, and the conditions of interplanetary space, 
through which the particles travel on their way to the earth. 
Although the discussion of these interesting questions lies beyond 
the scope of this book, I should like to mention two very significant 
results that have come from the study of many solar events. 

First, the particles produced in these events contmue to rain 
upon the earth for a period of hours after the disappearance of the 
flare from which they arose. Apparently the particles cannot escape 
directly from the solar system, but are trapped temporarily witlun 
it Second, the particles observed immediately after the flare often 


Cosmic Rays 

appear to approach the earth from a fairly well defined direction; 
toward the end of the event the particles appear to come from all 
directions. This is exactly what one would expect, because the 
"early" particles are those coming directly from the sun, whereas 
the "late" particles have bounced back and forth several times in 
space before reaching the earth. Even the early particles, of course, 
do not usually come precisely from the direction of the sun, because 
the magnetic fields present in interplanetary space bend their 

Forbush decreases 

While the solar-flare increases and the polar-cap absorption provide 
evidence for the production of high-energy particles by the sun, 
other observations show that solar activity also affects the flux of 
high-energy particles entering the solar system from outside. One 
striking effect is the worldwide decreases in cosmic-ray intensity 
that occur fairly frequently, especially during periods of high solar 
activity. The effect was first observed by the American physicist 
Scott E. Forbush, for whom it is named, in 1937. 

Forbush decreases are related to a number of geophysical 
effects the cause of which appears in every case to be a large solar 
eruption. Roughly one day after a solar eruption (the period varies 
from one event to the next), the earth experiences a magnetic 
"storm." Typically there is a short increase of the order of 1 part 
in 1,000 in the strength of the earth's magnetic field (sudden 
commencement), followed by a decrease (of the order of several 
parts in 1,000) that continues for a few hours (main phase). The 
magnetic field then slowly recovers its original strength over a 
period of days. 

Many, though not all, magnetic storms are accompanied by 
Forbush decreases. When they are, the cosmic-ray intensity begins 
to decrease at the same time as the magnetic field, reaching a 
minimum a few per cent below normal. And when the magnetic 

Cosmic rays and the sun 

field subsequently recovers, the cosmic-ray intensity also returns 
to its normal value (Fig. 14-2). Since Forbush decreases are a 
world-wide phenomenon, whatever agency produces them must 
affect both the low-energy primary particles (which can reach the 
earth's atmosphere only near the magnetic poles) and the high- 
energy particles (which can arrive at all latitudes). 

Simultaneously with the magnetic storms, there often occur 
large-scale changes in the outer radiation belt. From var.ous 





o +2 

[injv-i n iVk,ru 


r o 









c -4 



— L_ 



§ 100 


m 60 

'2 20 

i i i — r— i — I — i i i I — ' r~ ' 




1 1 1 1 1 1 1 — 1 1 1 ' 



14 15 16 17 

9 10 11 12 13 
Days, February 1958 

FiK 14-2 Typical Forbush decrease. The lower curve represents the hori- 
zontal component Afi of the magnetic field measured, from an arbitrary zero 
in units of 10- gauss. The upper curve shows the of 
cosmic-ray intensity (measured by a neutron monitor in Chicago) as « per- 
centage of the normal intensity. Shown on the horizontal axis is the universal 
time, ta days. (From F. Bachelet. P. Balata, A. M. Contort*, and G. Mann,. 
II Nuovo Cimento, vol. 16, p. 292, 1960.) 


Cosmic Rays 

satellite observations it appears that the radiation intensity drops 
sharply at first, particularly in the far region of the outer belt. At 
the same time, widespread aurorae appear. After a day or two, the 
outer belt begins to recover. The population of particles overshoots 
the prestorm level and then decreases gradually. After a few weeks 
the outer belt has returned to its normal state. 

A tentative picture of geomagnetic storms, Forbush 
decreases, and related effects 

A detailed explanation of all these diverse but clearly interrelated 
effects is still lacking. There is hardly any question that, at. the 
moment of a solar eruption, somelliing starts out from the sun and 
travels through space at such a speed as to reach the earth in about 
one day. This something is capable of producing magnetic storms, 
of partially shielding the earth from the oncoming cosmic-ray flux, 
of causing auroral displays, and of discharging the outer radiation 
belt. The same something also carries with it the energy needed to 
rebuild the outer belt. 

From a number of arguments based on astronomical observa- 
tions, on theoretical studies, and, more recently, on direct measure- 
ments by means of space vehicles, the following tentative picture 
has emerged. Out of the sun, or perhaps, more precisely, out of the 
solar corona, comes a continuous stream of diluted, highly ionized 
gas, or plasma, consisting chiefly of hydrogen. As a consequence, 
there exists in interplanetary space a more or less steady solar wind 
of plasma. Near the earth this wind has a speed of several hundred 
kilometers per second and a density of several particles per cubic 
centimeter. (Incidentally the experimental study of interplanetary 
plasma, under the direction of Herbert Bridge, is now one of the 
major activities of the MIT cosmic-ray group.) 

The "explosions" responsible for the solar flares occur at the 
base of the corona, in the region immediately above the photosphere, 
the visible disk of the sun. A sizable part of the energy released by 

Cosmic rays and Ihe sun 


an explosion appears as additional kinetic energy in the com- 
paratively dense plasma at the base of the corona. As a result, a 
cloud of plasma is driven through the corona into interplanetary 
space at a speed of roughly 1,000 kilometers per second. As it 
pushes against the interplanetary plasma (that is, the solar wind), 
which is also moving away from the sun but at a slower speed, the 
cloud produces a shock wave. According to this picture, the shock 
wave is the something previously referred to. 

Presumably upon hitting the earth's magnetic field, the shock 
wave at first compresses the lines of force, thus temporarily increas- 
ing the strength of the magnetic field (sudden commencement). 
Moreover, it may put kinks or wiggles in the distant lines of force 
(where the field is weaker, hence more easily disturbed), thus 
affording a possibility of escape to the trapped particles of the outer 


The main phase of the magnetic storm may result from the 
injection of solar plasma, by the shock wave, into the magneiosphere 
— the region in which the earth's magnetic field predominates 
over the interplanetary field. Alternatively, the shock wave may 
"heat" the plasma already present in the magnetosphere, that is, 
it may increase the individual energies of the plasma particles. 

In either case, the plasma pressure would increase, push the 
lines of force apart, and thereby decrease the field strength. This 
injection, or heating of the plasma in the magnetosphere, is 
probably closely related to the recovery of the outer Van Allen belt. 
Indeed, the plasma is thought to consist of ionized hydrogen, 
that is, of protons and electrons — the very particles found in the 
Van Allen belt. True, the energy of the plasma particles is, on the 
average, much lower than that of the particles detected by the 
counters used to study the Van Allen belt. But then the plasma 
particles and the Van Allen particles in the outer belt might be just 
the low-energy and the high-energy portions of a single particle 

You will recall that both Singer and Van Allen had suggested 
that particles of solar origin might be injected into the mag- 


Cosmic Fays 

nctosphere. Singer had in mind the low-energy plasma particles 
required to explain the main phase of the magnetic storms; Van 
Allen was referring to the particles of much higher energy that his 
counters had revealed. Together the two suggestions form a picture 
not very different from that favored by the scientific community 
at the time of this writing — although what is injected into the 
magnetosphere might be energy, not particles. 

Before leaving this subject, I must add a remark to avoid 
possible misunderstandings. Singer had pictured the main phase 
of geomagnetic storms as due to the electric currents resulting from 
the drift of trapped particles around the earth. The explanation 
given here ascribes the same magnetic disturbance to a pressure 
effect of the trapped plasma pushing the lines of force apart. The 
two points of view do not represent two different assumptions, but 
are two different ways of describing the same physical situation. 
Unfortunately, any attempt to explain why this is so would lead us 
into the thorny field of magneto-fluid dynamics. 

But what about Forbush decreases? The sun has a general 
magnetic field of several gauss; much stronger local magnetic fields 
usually exist in the disturbed regions where solar flares occur. The 
solar magnetic field induces electric currents in the ionized gases 
flowing out of the sun. Because these gases behave as nearly perfect 
electrical conductors, the currents continue to circulate for a long 
time in the gases as the latter move away from the sun. These 
currents, in turn, produce magnetic fields. Consequently, the solar 
wind in interplanetary space normally has a magnetic field, and the 
plasma immediately behind the shock-wave front resulting from a 
solar flare has an even stronger field. This field tends to keep 
cosmic-ray particles away, thus producing Forbush decreases. 

If the tentative picture is correct, Forbush decreases should 
occur not only in the vicinity of the earth but throughout inter- 
planetary space. This expectation has been confirmed by deep- 
space probes. For example, Pioneer V in 1960 detected a Forbush 
decrease at a distance of 5 million kilometers from the earth. At the 
same time, the magnetometer aboard Pioneer V (which had been 

Cosmic rays and the sun 


recording magnetic fields of the order of a few times 10" 5 gauss in 
interplanetary space) detected a large increase in the field strength 
(to about 4 X 10~ 4 gauss). The decrease in the cosmic-ray flux and 
the increase of the magnetic field strength probably marked the 
passage of a shock front. Shortly thereafter, the earth experienced 
a severe magnetic storm, presumably caused by the same shock 


Curiously enough, aurorae, although the first effect to be noted 
by scientists, are still the least understood among the various 
aspects of the sun-earth relationship. Aurorae appear as a lumi- 
nescence in the upper layers of the atmosphere. They usually occur 
at high geomagnetic latitudes and at altitudes ranging from about 
one hundred to several hundred kilometers. Instruments flown into 
the auroral regions by means of rockets (by the Iowa group) have 
detected large fluxes of electrons with energies between 10 and 20 
keV and protons with energies greater than about 70 keV (the 
instruments were not sensitive to particles of lower energy). 

On several occasions when artificial satellites were monitoring 
the Van Allen belt, aurorae were observed to occur simultaneously 
with the discharge of the outer portion of the belt. This observation 
seemed to suggest that aurorae might be due to particles escaping 
from the trapping region along the magnetic lines of force and into 
the atmosphere below. But further investigations have not sup- 
ported this view, and to date no other convincing explanation has 
been offered. 

The 11 -year cycle 

Solar scientists are well acquainted with the fact that the general 
pattern of solar activity, as evidenced by the number of sun spots, 
the frequency of solar flares, and other events, follows an 11-year 
cycle. When accurate records of the cosmic-ray intensity measured 
over long periods of time near sea level became available, and when 
the results of extensive balloon observations began to accumulate, 


Cosmic Rays 

it was found that the flux of cosmic rays also changes systematically 
during the 11-year solar cycle. 

For example, the counting rates of M-meson detectors near sea 
level are about 6 per cent lower when the solar activity is at a 
maximum than it is when the solar activity is at a minimum. More 
spectacular are the results obtained with balloon-borne detectors 
flown to higli altitudes near the magnetic poles, where the cosmic- 
ray flux is found to be several times smaller during maximum solar 








per cm' per 







c 200 









100 200 300 

Atmospheric depth, g/cm J 


Fig. 14-3 Intensity of cosmic rays as a function of atmospheric depth, 
measured at a geomagnetic latitude of 88° N at the time of minimum solar 
activity in 1954 and at the time of maximum solar activity in 1958. The 
instrument used was an electroscope; the vertical scale gives the number of 
ion pairs per second produced by cosmic rays in 1 cm' of air at standard temper- 
ature and pressure. The horizontal scale gives the atmospheric depth, in grams 
per square centimeter. (From H. V. Neher, Nature, vol. 184, p. 423, 1959.) 

Cosmic rays and the sun 


activity. This is shown by the curves in Fig. 14-3, which represents 
some of the extensive measurements made by H. Victor Neher of 
the California Institute of Technology. 

As I have already mentioned, only primary particles of several 
BeV or greater energy make their effects felt near sea level. On the 
other hand, primary particles of energies as low as 100 MeV can 
reach a balloon-borne instrument flown at high magnetic latitudes. 
Clearly, then, the agency responsible for the 11-year changes in 
cosmic-ray intensity has a much stronger effect on particles of com- 
paratively low energy than on those of very high energy. It seems 
very unlikely that the intensity variations of cosmic rays during 
the solar cycle are the result of changes in the emission of cosmic 
rays from the sun. For if th&y were the result of such changes, 
the greatest production of cosmic rays would occur at times of 
least activity. But, as we have seen before, the acceleration of 
particles is related to increased solar activity. 

A more appealing idea is that solar activity modulates the 
cosmic-ray flux. In agreement with the tentative picture that I 
presented in the discussion of Forbush decreases, we now believe 
that the magnetic field found in interplanetary space is, so to 
speak, carried materially away from the sun by the outflowing 
plasma. The interplanetary magnetic field acts as a partial screen 
against cosmic-ray particles entering the solar system from the 
outside. When the solar activity increases, more clouds are ejected 
from regions of the sun where strong local fields exist. As a conse- 
quence, the average magnetic field throughout interplanetary space 
becomes stronger, the screen becomes more opaque. 

Although this general interpretation seems quite plausible, no 
one can claim to understand in detail what actually happens. An 
interesting question is why the modulation mechanism responsible 
for Forbush decreases is about equally effective for high- and low- 
energy cosmic-ray particles, whereas the modulation mechanism 
responsible for the 11-year intensity changes acts much more 
strongly upon particles of lower energy. 

The origin of 

cosmic rays: an 

unsolved problem 


Half a century after the discovery of cosmic rays the problem of 
their origin is still unsolved. We do not know for certain where 
cosmic rays come from. We do not know for certain how they 
acquire their tremendous energies. 

Cosmic rays from the sun 

From the solar flare effects we do know that occasionally the sun 
accelerates particles up to energies of several BeV and more fre- 
quently up to energies of several hundred MeV. We know this 
acceleration occurs at times of high solar activity, when great 
eruptions take place and large masses of ionized gases shoot out 
from the sun into interplanetary space. We also know that magnetic 



Cosmic Rays 

disturbances accompany these eruptions, and we strongly suspect 
the fast-changing magnetic fields as the agency through which 
some of the protons in the solar atmosphere acquire high energies. 
We do not know in detail how this happens; the process may 
vaguely resemble the processes occurring in some high-energy 
particle accelerators, such as the betatron. 1 

Since the sun produces cosmic rays, why not assume that all 
cosmic rays come from the sun? In other words, why not assume 
that the sun emits high-energy particles continuously, thereby 
producing the normal cosmic-ray flux observed on the earth, and 
that the solar flare effects are just temporary increases in its normal 

The main argument against this assumption rests on the uni- 
form intensity of the radiation at all hours of the day and night. 
Cosmic rays come not only from the direction of the sun but from 
everywhere in the sky. 

The argument, however, is not as clear-cut as it may seem. The 
terrestrial magnetic field bends the trajectories of charged particles 
coming from the sun, making it possible for some of them to reach 
the night side of the earth (Chap. 5). Moreover, scientists have for 
some time suspected the existence of weak magnetic fields through- 
out the solar system. As noted in the preceding chapter, direct 
observations by space probes have confirmed their suspicions. 
Interplanetary magnetic fields will also scatter charged particles 
about, and these would not appear to come from the sun even if 
they originated there. 

Nonetheless, if cosmic-ray particles actually came from the 
sun, neither the earth's magnetic field nor the interplanetary mag- 
netic field could account for their almost perfectly uniform flow. 
This conclusion applies to particles with energies of the order of 
10 BeV, which form the bulk of the observed radiation. It applies 
even more forcefully to the particles of much higher energy 

1 The betatron is used to accelerate electrons to energies in the range of 
hundreds of MeV. The accelerating force is an electric field induced by a 
time-varying magnetic field. 

The origin of cosmic rays: an unsolved' problem 


responsible for air showers. A proton of 10" eV, for example, has a 
magnetic rigidity (see pages 56 and 246) 


BR = ~q = 3 X 10 12 gauss -cm 

Since the interplanetary field B is of the order of several times 
10~ s gauss — say, 3 X 10" 4 gauss — the radius of curvature R of 
this proton in interplanetary space is somewhere around 10" cm, 
which is about 6,500 times the distance from the earth to the sun 
(1.5 X 10" cm). On the other hand, since the earth's magnetic 
field is of the order of 0.5 gauss, the radius of curvature in the 
vicinity of the earth is approximately 6 X 10 12 cm, or 10,000 times 
the radius of the earth (6.38 X 10 8 cm). 

Thus protons with energies of 10" eV or more do not undergo 
any appreciable deflection either in the interplanetary or in the 
terrestrial magnetic field. If they were produced by the sun, they 
would come only from the direction of the sun. On the contrary, 
as the air shower experiments have shown, there is no preferred 
direction of arrival whatsoever. 

Cosmic rays from the stars 

If the sun does not contribute a major fraction of the observed 
cosmic radiation, then where does one look for the main source of 
cosmic rays? There are, of course, billions upon billions of stars 
in the universe, and one might think that what reaches the earth is 
the combined cosmic-ray emission of all these stars. Considering 
the relative nearness or the sun, however, if the average t -■ Dro- 
duced the same number of cosmic rays as the sun does, the total 
flux upon the earth of cosmic-ray particles from the sun would be 
millions of times greater than that from all other stars together. 
This would still mean that practicall; -M r ^mic. rays had to come 
from the sun. And they simply do not. 

Of course, there may be stars with special properties that make 
them particularly effective sources of cosmic rays. Conceivably, 



Cosmic Rays 

relatively small numbers of such stars may supply the whole of the 
observed cosmic-ray flux. 

Supernovae have also figured prominently in the speculations 
on possible cosmic-ray sources. These gigantic explosions of indi- 
vidual stars occur once every few hundred years in our galactic 
neighborhood (that is, within several thousand light years of the 
solar system). An explosion releases an amount of energy equiva- 
lent, perhaps, to the total mass of the sun. 

Nobody knows for sure why the gradual process of nuclear 
fusion — which supplies the energy continuously radiated into 
space by ordinary stars — sometimes degenerates into an explosive 
release of energy. But most astrophysicists agree on the following 
broad picture. Nuclear fusion occurs predominently near the 
centers of stars, where the temperature is highest. Eventually, the 
nuclear furnace in the innermost region runs out of fuel; this 
means. that all primeval hydrogen has fused into helium and helium 
has fused into progressively heavier nuclei, until the center of the 
star is practically all iron, the most stable of all nuclei. 

At this point the gravitational forces, no longer balanced by 
the radiation pressure due to the continuous generation of energy, 
will cause the star to collapse. The implosion produces a rapid 
increase of temperature that is due to the transformation of gravi- 
tational energy into heat. The hydrogen and the other light nuclei, 
which arc still present in the peripheral regions of the star, now 
begin to fuse. And since the rate of fusion increases rapidly with 
temperature, most of the remaining nuclear fuel burns up nearly 
at once, and a tremendous explosion results. 1 This catastrophic 
process is held responsible for the synthesis of the heavier nuclei. 

1 Not every star blows up at the end of its life. It has been shown that a 
supernova explosion can occur only if the star has a mass greater than 1 14 
times that of the sun. Other properties (such as the speed of rotation and the 
chemical composition) may play a role. Note that the gravitational collapse 
will acquire catastrophic proportions only if the heat energy released in the 
interior of the star escapes fast enough; for otherwise it will create a pressure 
that slows down the process. It is believed that the energy escapes in the 
form of neutrinos. 

The origin of cosmic rays: an unsolved problem 


It has been suggested that supernovae also give rise to high-energy 
particles — that is, cosmic rays, and various ingenious ideas have 
been put forward to explain how this could happen. 

Most scientists believe, however, that cosmic-ray particles are 
accelerated by electromagnetic processes more or less like those at 
work in solar flares. Astronomers have observed stars in which 
eruptions resembling solar flares occur with great frequency. They 
have also observed double stars that possess large magnetic mo- 
ments and revolve rapidly around a common center of mass, thus 
producing strong and varying magnetic fields. Finally, there are 
some variable stars whose exceedingly strong magnetic fields 
appear to reverse direction every Tew days. All of these are possible 
sources of cosmic rays. 

Cosmic rays and radio noise 

Several scientists, and particularly the Russian astrophysicist 
V. L. Ginzburg, have pointed out that radioastronomical observa- 
tions may offer important clues to the origin of cosmic rays. Their 
line of reasoning runs as follows. If an electromagnetic field accel- 
erates the protons and heavier nuclei found in cosmic radiation, it 
presumably accelerates electrons as well. The magnetic field, which 
is necessarily present in the region where the particles are accel- 
erated, bends the trajectories of the electrons and obliges them to 
spiral around the lines of force. 

Since the motion along a curved path is an accelerated motion, 
the spiraling electrons must lose energy by emitting electro- 
magnetic radiation (see Chap. 6). This effect is well known to all 
physicists working with high-energy accelerators. It is responsible 
for the visible and ultraviolet light emitted by fast electrons as they 
circle around and around in a synchrotron, and it is known as 
synchrotron radiation. The same name is used to describe the 
radiation produced by the magnetic tending of fast-electron tra- 
jectories in space. 


Cosmic Rays 

In principle, of course, all charged particles moving in a 
magnetic field emit synchrotron radiation But the intensity of 
this radiation is critically dependent on .nass of the particles. 
Lighter particles, which are more easily deflected by a magnetic 
field and which move with higher speeds, radiate much more 
strongly than heavier particles. In fact, the synchrotron radiation 
of protons and other nuclei is completely negligible under any 
conditions, either in the laboratory or, from what we know, in the 

Synchrotron radiation by electrons, on the other hand, is 
often a major effect. Its relative intensity in the various parts of 
the electromagnetic spectrum depends on the strength of the mag- 
netic field and on the energy distribution of the electrons. Syn- 
chrotron radiation in the spectral region corresponding to radio 
waves contributes to the radio noise arriving from outer space. 
Radio astronomers have learned how to distinguish this type of 
radio noise from types arising from different sources. 

What I have just said makes it appear likely that the sources 
of cosmic rays and the sources of synchrotron radiation are closely 
related. Our nearest star, the sun, is a natural subject for a direct 
experimental test of this view. A number of observations have 
shown that solar flares that give rise to high-energy protons are 
almost consistently accompanied by the outbursts of radio noise 
known as type I V bursts. Astrophysicists agree that type IV radio 
bursts are the result of synchrotron radiation by high-energy 
electrons spiraling around the magnetic field lines in the solar 

Other celestial objects that are strong sources of radio noises 
have been singled out as likely sources of cosmic rays; among them, 
in particular, are the gas clouds left behind by supernova explo- 
sions. Five or six such remnants of outbursts recorded in historical 
documents during the last 2,000 years or so are now visible in the 
sky. The best known is the Crab nebula, a gas cloud about six light 
years in diameter and about 4,000 light years away, the remnant 
of a supernova explosion observed by Chinese astronomers in 

The origin of cosmic rays: an unsolved problem 


aj>. 1054 (a light year — the distance traveled by light in a year — 
is approximately 10 18 cm). The gas within the cloud is still in a 
state of violent agitation; the motions are clearly discernible from 
a comparison of pictures of the nebula taken a few years apart. 
Much of the visible light emitted by the Crab nebula shows a 
continuous spectrum. For many years this was a puzzle, because 
such dilute gases usually radiate at several sharply defined fre- 
quencies. It was later found that this light is also strongly polar- 
ized. The Russian astrophysicist I. S. Shklovsky then pointed out, 
in 1953, that the only sensible explanation of both the continuous 
spectrum and the polarization lies in the assumption that the light 
is the product of a synchrotron radiation. In the meantime, radio 
astronomers had also identified the Crab nebula as a strong source 
of radio noises. Astrophysicists now attribute the radio emission, 
as well as the light emission, to a synchrotron radiation. 

The great intensity of the synchrotron radiation from the Crab 
nebula and the appearance of much of this radiation in the visible 
region of the spectrum indicate the presence of large numbers of 
electrons, with energies extending up to hundreds of BeV, circling 
around in relatively strong magnetic fields. Presumably, the mecha- 
nism responsible for the acceleration of the electrons also accel- 
erates protons and whatever other nuclei may be present in the 
nebula. These particles may escape from the Crab nebula into 
interstellar space. If they do, they may, together with particles 
originating in the same manner from the remains of the other 
supernovae, contribute a major portion of the observed cosmic 

The Fermi acceleration mechanism 

Protons and nuclei of the heavier elements are found everywhere in 
the universe. The problem is not to explain their presence in the 
cosmic radiation, but rather to figure out how they might have 
acquired their tremendous energies. So far, we have considered 


Cosmic Rays 

acceleration mechanisms operating within restricted regions of 
space, sucli as stellar atmospheres or remnants of supernovae. 
However, it is also possible that cosmic-ray particles are accel- 
erated gradually, while traveling through space, at the expense of 
electromagnetic fields present in the interstellar medium. In 1933, 
W. F. G. Swann, director of the Bartol Research Foundation, 
suggested that the galaxy as a whole might act as a gigantic 
accelerator of cosmic rays. The principle of the acceleration mecha- 
nism proposed by Swann was, in essence, the one later applied in 
the development of the betatron. 

In 1951, Enrico Fermi suggested a more realistic mechanism, 
based on the general picture of interstellar space that was beginning 
to emerge from astronomical observations. According to this pic- 
ture, enormous clouds of ionized gas, mostly hydrogen, wander 
through space. The clouds contain magnetic fields, just as do the 
gas clouds ejected from the sun. Fermi pointed out that energy is 
exchanged between these wandering clouds and fast-moving indi- 
vidual particles and that, on the average, the individual particles 
gain energy in the exchange. 

To understand bow this happens, consider the following model, 
which, although oversimplified, brings out the essential features of 
the process. Assume the clouds are more or less rigid bodies and 
that the space between them is free of magnetic fields. Then a 
particle will travel in a straight path until it strikes a cloud. Once 
inside a cloud, it describes a complicated path, and it eventually 
emerges will) practically no "memory" of its original direction of 
motion. Therefore, according to the model, particles scatter back 
and forth between clouds. 

When a particle collides with a cloud coming toward it, it 
bounces off with increased energy ; when it collides with a receding 
cloud, it loses energy. (Any tennis player knows he can increase or 
decrease the speed of a ball that hits his racket by moving the 
racket forward or backward.) On the average, the particle obtains a 
small net gain of energy, largely because head-on collisons occur 
more frequently. Fermi thought of this mechanism as operating 

The origin of cosmic rays: an unsolved problem 


within our galaxy. A similar mechanism may be at work in a more 
limited region of space, such as the Crab nebula or the atmospheres 
of stars. Also, it may be at work, on a more grandiose scale, in the 
regions of space between galaxies. 

Cosmic rays in the solar system, in the galaxy, 
and beyond 

The flux of energy reaching the earth in the form of cosmic rays is 
approximately the same as that received in the form of light from 
all stars, excluding the sun. The question of how the cosmic-ray 
flux found near the eartli compares with that in distant parts of 
the universe is not as easily answered. The problem is much more 
difficult in the case of cosmic rays than in the case of star light, 
because light travels in straight lines and cosmic-ray particles, 
being electrically charged, are deflected by magnetic fields. Mag- 
netic deflection may greatly retard the escape of cosmic-ray 
particles from the regions of space in which they are produced, and 
thus cause local concentrations in the particle population. In fact, 
as the discovery of the radiation bell around the earth has shown, 
under appropriate circumstances charged particles may remain 
magnetically trapped for long periods of time. 

If very effective trapping fields were present in the solar sys- 
tem, a relatively small supply of high-energy particles from the sun 
would suffice to maintain the observed flux of cosmic rays. More- 
over, since most cosmic rays would reach the earth after bouncing 
back and forth many times in the surrounding space, they would 
appear to come from every direction, rather than only from the sun. 
Then the sun, and presumably other stars as well, would be sur- 
rounded by an "atmosphere" of cosmic rays, whereas the cosmic- 
ray flux in interstellar space would be negligible. But, as we have 
seen, the magnetic fields in the solar system are not strong enough 
for this purpose. (Apparently they are only able to store solar- 
flare particles of energies up to several BeV for a few hours or days.) 


Cosmic Rays 

This is why physicists were compelled to look beyond the solar 
system in their search for the source of cosmic rays. Clearly, if 
cosmic rays do not originate within the solar system, they must fill 
a large volume of interstellar space around the sun. 

Stars and interstellar matter are not distributed uniformly 
throughout the universe but are condensed in galaxies. Our own 
galaxy contains about 100 billion (10") stars. Most of these, 
particularly most of the younger, more active stars, occupy a flat 
volume shaped roughly like a grindstone with a bulge at the 
middle. The diameter of the galactic disk is approximately 100,000 
light years, and its thickness is a few thousand light years. Much 
of the interstellar gas and dust is also condensed in this volume. 
The disk, however, is surrounded by a halo, roughly spherical 
in shape, formed by old stars and very dilute gas (Fig. 15-1). 

It is natural to assume that cosmic rays are produced in 
galaxies rather than in the nearly empty space between galaxies. 
If this is so, then most of the observed cosmic radiation should 

1,000 light years 

5,000 light years 

Fig. 15-1 The structure of the Milky Way galaxy, represented schematically. 
The galactic disk has an average thickness of about 1,000 light years (10" cm); 
the central bulge in the disk is about 5,000 light years thick. The diameter 
of the disk is about 100,000 light years (10" cm), and the disk is surrounded 
by a nearly spherical halo of the same diameter. The solar system is located 
near the median plane of the disk, about two-thirds of the way from the center. 

The origin of cosmic rays: an unsolved problem 

come from our own galaxy. Other galaxies, because of their great 
distance, should not contribute more than a small fraction of the 
cosmic-ray flux, just as they do not contribute more than a small 
fraction of the light flux in the night sky. 

The solar system is located near the median plane of the 
galaxy, about two-thirds of the way from the center. The dis- 
tribution of galactic stars and galactic interstellar matter with 
respect to the earth is therefore very uneven. This, of course, 
explains the appearance of the great concentration of stars in the 
Milky Way. Presumably, the sources of cosmic rays are also dis- 
tributed very unevenly. If there were no magnetic fields in the 
galaxy, the intensity of cosmic radiation reaching the earth from 
different directions would vary. One would expect to find most 
cosmic rays coming from the center of the galaxy. Yet the intensity 
of cosmic rays from different parts of the sky does not vary by 
more than a fraction of 1 per cent. Thus, if cosmic rays are of 
galactic origin, there must be magnetic fields in the galaxy capable 
of producing a random distribution of cosmic radiation in space. 
I have already mentioned the belief that the wandering clouds 
of ionized gases contain magnetic fields. The crude picture repre- 
senting cosmic-ray particles as bouncing back and forth between 
magnetic clouds would explain not only their acceleration but also 
the random distribution of their directions of motion. Be that as it 
may, it is quite likely that magnetic fields keep cosmic-ray par- 
ticles trapped in the galactic volume for long periods of time 
(compared to the time required to escape along straight lines). 
Our galaxy, then, and presumably other galaxies as well would 
have their own cosmic-ray populations, distributed more or less 
uniformly throughout their volumes, while in the space between 
galaxies the density of cosmic rays would be almost negligible. If 
this view is correct, even intermittent sources of cosmic rays within 
each galaxy (for example, supernovae) might maintain a fairly 
steady cosmic-ray flux. 

Whether cosmic rays are trapped within the galactic disk or 
within the galactic halo, and how long the actual trapping time is, 


Cosmic Rays 

are still subjects of debate. One important clue is the composition 
of cosmic rays. The very fact that the cosmic radiation contains 
nuclei of heavy elements sets an upper limit to the amount of 
matter traversed by cosmic rays from the moment they are pro- 
duced to the moment they are detected. For instance, interstellar 
gas in the galactic disk has an average density of about one atom 
of hydrogen per cubic centimeter. Considering the known proba- 
bility for collisions between nuclei of hydrogen and carbon, a 
carbon nucleus will travel for about 4 million years in the galactic 
disk before colliding with a hydrogen nucleus of interstellar gas. 
Such a collision would destroy the carbon nucleus. Consequently, 
since carbon nuclei are found in cosmic radiation, the radiation 
cannot remain confined in the galactic disk for more than a few 
million years on the average. 

Another important clue is the small number of high-energy 
electrons in primary cosmic radiation. This may well result from 
the fact that electrons moving in a magnetic field emit synchrotron 
radiation and thereby lose energy fairly rapidly, whereas protons 
and heavier nuclei do not undergo a corresponding energy loss. 
Processes such as those occurring in solar flares may not be able to 
accelerate electrons beyond a certain limit because of the fast rate 
at which synchrotron radiation robs electrons of their energy. Also, 
if cosmic rays are magnetically trapped for long periods of time, 
the energy loss by synchrotron radiation may drastically reduce 
the flux of high-energy electrons. 

Until a few years ago it was generally believed that all cosmic 
rays arriving at the earth were produced in our own galaxy. This 
belief has been seriously shaken by the results of air-shower experi- 
ments that have gradually pushed the upper limit of the cosmic-ray 
spectrum to higher and higher energies. The occurrence of primary 
cosmic rays with energies at least as great as 6 X 10 19 eV has been 
established. If these particles are protons, their magnetic rigidity is 

6 * 10 19 
BR = 3 q = 2 X 10" gauss- cm 

Now, several pieces of evidence indicate that the magnetic field 

The origin of cosmic rays: an unsolved problem 


in the galactic disk has an average value of about 5 X 10~ 6 gauss. 
Presumably the magnetic field in the halo is no stronger. Thus, 
both in the disk and in the halo the radius of curvature of the 
particles considered here is at least 4 X 10" cm. This radius is 
about 40 times greater than the thickness of the galactic disk 
(10" cm) and about as large as the radius of the galactic halo 

(5 X 10" cm). 

It does not seem possible for particles so little affected by the 
galactic magnetic fields to remain trapped in the galactic disk or 
even in the galactic halo for any great length of time. If they do 
not remain trapped, then one must conclude that they are to be 
found in the space between galaxies, as well as in the space within 
galaxies, and that most of those observed on earth come from the 
space beyond our galaxy. 1 

A word of caution is necessary. I have assumed that the high- 
energy particles responsible for air showers are protons. This is not 
certain. If they are heavier nuclei, they have, because of their 
greater electric charge, a smaller magnetic rigidity than protons 
of the same energy (see pages 57 and 246). For example, in a mag- 
netic field of 5 X 10~ 6 gauss, an iron nucleus (Z = 26) of 6 X 10 19 
eV energy has a radius of curvature of little more than 10" cm, which 
is much less than the radius of the galactic halo. These particles, 
therefore, could conceivably remain trapped in the galactic halo, 
if not in the galactic disk. 

On the other hand, there is no indication that the cosmic-ray 
spectrum breaks off at 6 X 10 19 eV. Any new experimental data 
showing that the spectrum actually continues well beyond this 
energy, or any new results showing that the high-energy cosmic-ray 
particles are protons rather than heavier nuclei, could lead to only 
one conclusion: these particles are of extragalactic origin. 

' This conclusion is not inconsistent with the view that the total cosmic- 
ray population in the space between galaxies is small compared to that within 
the galaxies. The evidence for an extragalactic origin applies only to cosmic-ray 
particles of very great energy, and these form only a minute fraction of the total 
cosmic-ray flux. As Fig. 11-6 shows, for example, only one primary particle 
in 10" has an energy greater than 10 ls eV. 



Cosmic Fays 


The last pages of this book are written in August, 1962. A few days 
ago it was the fiftieth anniversary of Hess's flight, with which my 
story began. The half century covered by this story has been a 
revolutionary period for science. And cosmic rays, as I have tried 
to show, have played a major role in the developments that have so 
greatly enlarged the horizon of our knowledge. Without the wholly 
unexpected facts and without the tantalizing clues that came to 
light through the study of cosmic rays, high-energy physics might 
still be in its infancy. Indeed, physicists might not yet have dis- 
covered mesons and all the other particles that have been their 
major concern during the last decade. And it is certainly not a mere 
coincidence that the first scientific discoveries of the space age — 
including the discovery of the Van Allen radiation belt — were 
made by cosmic-ray physicists. 

It is particularly appropriate at this time to pause and look 
back on the history of cosmic rays, not so much because the fiftieth 
anniversary of their discovery calls for some sort of celebration, but 
because, curiously enough, the anniversary comes at a critical 
moment for cosmic-ray physicists, if not for cosmic-ray physics 
itself. The interest in cosmic rays is certainly not waning; on the 
contrary, it is steadily growing. But cosmic-ray research has 
become such an integral part of many different scientific endeavors 
that it has almost ceased to exist as a separate and distinct branch 
of science. The "cosmic-ray physicist," as a specialist, is becoming 
a figure of the past, while the nuclear physicist, the geophysicist, 
the astrophysicist, and the cosmologist are turning more and more 
to the study of cosmic rays for information of vital importance to 
the solution of their problems. It is quite possible that future his- 
torians of science will close the chapter on cosmic rays with the 
fiftieth anniversary of Hess's discovery. However, they will 
undoubtedly note that in renouncing its individuality and merging 
with the main stream of science, cosmic-ray research continued to 
perform a vital role in advancing man's understanding of the 
physical world. 

Mass per unit area 



The concept of mass per unit area appears sufficiently often in this 
book that a few words of explanation may be useful. When com- 
paring radiation absorbers of different substances, it becomes neces- 
sary to consider the density (mass per unit volume), as well as the 
thickness, of the absorbers. Obviously, 1 meter of air will absorb 
much less radiation than 1 meter of water, because the density or 
one is so much smaller than that of the other (0.00129 g/cm' 
compared to 1.0 g/cm 3 at standard temperature and pressure). 
Thus, it is customary to define an absorber not by its geo- 
metrical thickness, but by the mass of a column of unit cross- 
sectional area (Fig. A-l). This quantity - the mass per unit 
area — is usually measured in grams per square centimeter 
(g/cm 2 ). For an absorber of constant density, the mass per unit 
area is just the product of its thickness and its density. For 
example, a layer of water 1 meter thick has a mass per unit area of 
100 g/cm s ; the same thickness of air has a mass per unit area of 
100 X 0.00129 = 0.129 g/cm s . At sea level, the earth's atmosphere 
has a mass per unit area (which in this instance is numerically 
equal to the atmospheric pressure) of 1,003 g/cm 2 . The mass per 
unit area of the atmosphere above a given level is known as the 
atmospheric depth. 



Cosmic Hays 

1 cm 

Fig. A-l Definition of mass per unit area of an absorber. Consider a prism 
of 1 cm by 1 cm in cross section cut from an absorber slab of thickness It. 
The mass of this prism is the mass per unit area of the absorber, in grams per 
square centimeter. The definition applies both to an absorber of uniform 
density and to an absorber made of layers of different density (such as the 
earth's atmosphere). In the former case, the mass of the prism is just its 
volume (1X/i= h) times the density d, and it is therefore equal to hd. 

Powers of ten 



In the treatment of very large and very small numbers in this book, 
I have assumed that the reader is more or less familiar with the 
numerical notation based on positive and negative powers of 10. 
The following notes, therefore, are intended only as a brief re- 
fresher for any reader who may find it useful. 

Numbers exactly equal to the products of n terms equal to 10 
are represented as the nth power of 10, where n is a positive 

10' = 10 

10 s = 10 X 10 = 100 

10» = 10 X 10 X 10 = 1,000 

10 8 = 1,000,000 = 1 million 

10" = 1,000,000,000 = 1 billion 

A number greater than 10 but not equal to a positive power 
of 10 is represented as the product of a number between 1 and 10 
and a power of 10. To take a specific instance, the approximate 
speed of light in a vacuum is 

1.86 X 10 5 = 186,000 miles per second, or 
3 X 10 6 = 300,000 kilometers per second, or 
3 X 10 8 = 300,000,000 meters per second, or 
3 X 10 10 = 30,000,000,000 centimeters per second 



Cosmic Rays 

A negative power of 10 is the reciprocal of the corresponding 
positive power: 

10- 1 = 1/10 = 0.1 

10-* = 1/10 1 = 1/100 = 0.01 

10-' = 1/10» = 1/1,000 = 0.001 

10-» = 1/10 6 = 1/1,000,000 = 0.000,001 = 1 millionth 

10-" = 1/10 9 = 1/1,000,000,000 = 0.000,000,001 - 1 billionth 

A number between and 1, but not equal to a negative power 
of 10, is represented as the product of a number between 1 and 10 
and a negative power of 10. To take a specific instance, the mass of 
an electron, in grams, is 

9.11 X 10-" = 

10 28 

= 0.000,000,000,000,000,000,000,000,000,911 gram 

Logarithmic scales 



In cosmic-ray physics, more perhaps than in other areas of physics, 
the quantities of interest vary over a wide range. An extreme 
example is the measured energy spectrum of primary cosmic rays 
(Fig. 11-6), where the energy varies from about 10» to about 10" 
e V _ that is, by a factor of 10 10 — and the corresponding intensity 
from a value of the order of 10~ l to a value of the order of 10"" 
particles per square centimeter per unit solid angle per second — 
that is, by a factor of the order or 10 18 . In these cases, the usual 
(linear) graph — where the distance from the origin is proportional 
to the magnitude of the number — does not afford a convenient 
method of representation. If, for example, a horizontal scale 10 cm 
long represented the energy range from to 10" eV, all measure- 
ments in the range, say, from 10» to 10" eV, would have to be 
crowded into a ridiculously small fraction of this scale, precisely 
into 10- 7 cm! (10" - 10« = 9.9 X 10"; 9.9 X 10"/10" = 9.9 X 
10-" « 10-"; 10 cm X 10~ 8 = 10" 7 cm). 

Thus, if one wishes to cover a large range of values in a single 
graph, one must use a scale that "shrinks" as the numbers increase. 
Such a scale is the logarillimic scale, like that used on the horizontal 
and vertical axes of Fig. 11-6. On a logarithmic scale, the distance 
from the origin represents the logarithm to the base 10 of the 



Cosmic Rays 

number, rather than the number itself. For example, the distance 
between 10" and 10" is the same as the distance between 10" 
and 10 19 (because log 10 10" - log 10 10' = 11 - 9 = 2, and 
log I0 10" - Iog.o 10" = 19 - 17 = 2). In general, the distance 
between two points is proportional to the ratio of the two corre- 
sponding numbers (100 in the example above) rather than to their 
arithmetical difference. 

Energy and 



The kinetic energy of a particle is the energy the particle possesses 
as a result of its motion. When a beam of particles (for example, 
electrons in a cathode-ray tube) hits a target, the kinetic energy of 
the particles is changed into heat; the amount of heat developed 
is proportional to the number of particles that hit the target 
multiplied by the average kinetic energy of the particles. According 
to classical (that is, nonrelativistic) mechanics, the kinetic energy 
E is related to the mass of the particle m and to its velocity v by 
the equation 

E =\mv* (D-D 

The equation shows that for a given velocity the kinetic energy is 
proportional to the mass; for example, an a particle, which is 
nearly 4 times heavier than a proton, has a kinetic energy nearly 4 
times that of a proton of the same velocity. The equation also 
shows that for a given mass the kinetic energy is proportional to 
the square or the velocity ; thus, if the velocity of a given particle 
is doubled, its kinetic energy becomes 4 times greater. 

Albert Einstein showed that Erj. (D-l), once believed to be 
rigorously correct, actually represents an approximate expression 



Cosmic Rays 

of the kinetic energy. The equation is most exact when the velocity 
of the particle is very small compared to the velocity of light 
(c = 300,000 km/sec), but it is grossly wrong for velocities com- 
parable to c. The correct equation for the kinetic energy, derived 
from Einstein's theory of relativity, reads 

E = 


\Vi - v 2 /c 2 l ) 


In Eq. (D-2) as well as in the classical equation (D-l), E is 
proportional to m. Moreover, one can prove mathematically that 
Eq. (D-2) reduces to Eq. (D-l) at the limit for v -» 0, as it should. 
But Eq. (D-2) shows that, when v is comparable to c, E increases 
at a faster rate than the square of the energy. Indeed, as v ap- 
proaches c, the squ are root Vl — v 2 /c 2 approaches zero and the 
ratio 1/v 1 — v 2 /c 2 approaches infinity, as does the kinetic energy 
E. In physical terms this result means that no material particle 
can ever travel at the velocity of light, simply because to give a 
particle an infinite amount of kinetic energy would require an 
infinite amount of work. 

The quantity mc 2 , which appears as a factor on the right-hand 
side of Eq. (D-2), is known as the rest energy of the particle. 
According to Einstein's equivalence principle which reads 

Energy = mass X c* (D-3) 

the rest energy mc- represents the energy "stored" in the par- 
ticle's mass. This is the energy released if the particle is annihilated; 
thus, when an electron and a positron annihilate each other (Chap. 
6), two photons are produced with a total energy equal to c 2 times 
their combined mass. Writing Eq. (D-3) in the equivalent form 

ma83 = «?m (D^ 

we see that the equivalence principle assigns a definite mass (that 
is, a definite mechanical inertia) to energy. This form of the 
equivalence principle has also been verified experimentally. When, 
for example, two protons and two neutrons combine to form a 
helium nucleus (or a particle; see Appendix G), a certain amount 
of energy AE is released. Direct measurement shows the a-particle 
mass to be slightly less than the combined original mass of the two 
protons and two neutrons. This mass defect is exactly equal to 
AE/c 2 . 

Energy and momentum — appendix D 


The upper curve in Fig. D-l provides a graphical representa- 
tion or Eq. (D-2). The ratio E/mc 2 of the kinetic energy to the rest 
energy is plotted on a logarithmic scale on the horizontal axis. The 
vertical (linear) scale represents the ratio v/c of the velocity of the 
particle to the velocity of light. The graph shows that the particle 
velocity approaches the velocity of light when E is substantially 
greater than mc' 1 (for example, v becomes greater than 0.95c for E 
greater than 2.2 mc 1 ). In this case, the particle is said to be rela- 
tivislic. Clearly, light particles (for example, electrons) become 
relativistic at a lower energy than do heavy particles (for example, 


The momentum p is another property of a moving particle. 
Unlike the kinetic energy, which is a scalar quantity, the momen- 
tum is a vector. The physical significance of the momentum lies 
in the second law of dynamics, which states that a force F acting 
on a given object during a lime t will increase the momentum of the 
object by an amount Ft. Thus, the initial momentum of the 
object p,, its final momentum p* and the force F arc related by 
the vector equation 

I* - p, = ¥1 (D- 5 ) 


1 — 1 — 1 1 1 1 1 II 

1 1 — i | i i 1 1 1 

1 — i — i | i i ii 


1— r-r , ■ ■ . . 



E / 


s /— 

s / mc 




. . .1 

. ..!.... 

i *■ — 

0.01 0.1 

1 10 


E/mc 2 (upper curve) or p/mc (lower curve) 

Fig. D-l The velocity of b particle r, expressed as a fraction of the velocity 
of light c, as a function of the kinetic energy E divided by the rest energy mc', 
and as a function of the momentum p divided by mc. 


Cosmic Rays 

Classical mechanics provides the following (approximate) 
expression for the momentum p of a particle of mass m moving with 
a velocity v: 

p = mv (D-6) 

Here the momentum is seen to be proportional to the mass of the 
particle and to vary as the first power of the velocity. The correct 
relativistic equation for the momentum is 

p-vra' < D - 7 > 

Like the kinetic energy, the momentum of a particle approaches 
infinity as the velocity approaches the velocity of light. The lower 
curve in Fig. D-l provides a graphical representation of the 
momentum-velocity relation. Plotted on the horizontal axis is the 
quantity pc/mc s (= p/mc), that is, the momentum multiplied by 
the velocity of light and divided by the rest energy of the particle. 
Plotted on the vertical axis is the ratio v/c. Comparison of the two 
curves in Fig. D-l brings out the fact that for relativistic particles 
the kinetic energy E is nearly equal to pc: 

E~pc (D-8) 

[One can reach the same conclusion analytically from Ivqs. (D-2) 
and (D-7), c onsidering that, when v is close to c, the ratio 
1/Vl — » 2 /c 2 is large compared to 1 ; therefore, Eq. (D-2) 

E « mcWl - «Vc 2 

On the same ass umption, c can be substituted for v in Eq. (D-7) to 
yield p = mc/vl — « 2 /c 5 ; therefore, pc ~ E.\ 
From Eq. (D-2) we obtain 

E + mc i = /-; ttt 

V 1 - » 2 /c ! 

and therefore 


= V\- v t /c 

E + mt* 

If the mass of the particle is zero, the term on the left-hand side 
of the equation vanishes. Thus \/l — w 2 /c 2 = and therefore 
v = c. We conclude, then, that a particle oj zero mass (such as a 
photon or a neutrino) must always travel al Hie velocity of light, 
whatever its energy. 

The electron volt 




The kinetic energy of the individual electrons in a cathode-ray 
beam depends on the potential difference (number of volts) be- 
tween the hot negative filament (cathode) and the positive anode. 
In fact, the kinetic energy is simply proportional to this potential 
difference and does not depend on the distance between the two 
electrodes or on the shape of either one. For example, a 20,000-volt 
potential between the cathode and anode of any cathode-ray tube 
will always produce electrons of the same kinetic energy, and the 
energy will always be twice that of electrons generated by a 
10,000-volt potential. 

This suggests a convenient unit for defining the energies of 
individual particles. An electron accelerated by a potential differ- 
ence of 20,000 volts has a kinetic energy of 20,000 electron volts (or 
20,000 eV for short); an electron accelerated by a potential 
difference of 10,000 volts has a kinetic energy of 10,000 eV; and so 
on. In other words, the basic unit of energy (the electron volt) is 
the energy acquired by an electron when it is accelerated by a 
potential difference of one volt. The multiples of this unit most 
commonly used are 

1 keV = 1,000 eV (1 thousand electron volts) 
1 MeV = 10* eV (1 million electron volts) 
1 BeV = 10« eV (1 billion electron volts) 


Cosmic Rays 

Thus, the statement that an a particle has an energy of 4.9 MeV 
means that its kinetic energy equals that of an electron accelerated 
by a potential difference of 4.9 million volts. 

Any electrically charged particle is accelerated by an electric 
potential difference. From general principles of mechanics, it 
follows that particles that have the same electric charge and are 
accelerated by the same potential difference acquire equal kinetic 
energies even if their masses are different. For instance, a proton 
accelerated by a potential difference of 10,000 volts will acquire the 
same kinetic energy (10,000 eV) as an electron accelerated by a 
potential difference of the same magnitude. On the other hand, 
particles of different charge accelerated by a given potential differ- 
ence acquire kinetic energies proportional to their charge. A 
potential difference of 10,000 volts (which accelerates a singly 
charged proton to a kinetic energy of 10,000 eV) will accelerate 
a helium nucleus (which contains two protons and therefore has 
two elementary charges) to a kinetic energy of 20,000 eV. 

The rest energy of a particle is also conveniently expressed in 
electron volts. For an electron, mc* = 0.51 MeV, and for a proton, 
mc* = 938 MeV. 

In the standard system of units, the unit of energy is the joule. 
One joule, which is approximately the energy required to raise a 
weight of 1.1 kilograms to a height of 10 cm, equals 6.24 X 10 18 eV. 

Computation of 
magnetic rigidity 



The radius R of the circle described by a charged particle moving 
in a uniform magnetic field, in the plane perpendicular to the lines 
of force (Chap. 5), can be obtained from the condition that the 
centrifugal force and the Lorentz force must balance. If m is the 
mass of the particle and its velocity, and if v is small compared to 
the velocity of light, the centrifugal force is mxr/R. If B is the 
strength of the magnetic field and if the electric charge of the 
particle is Z limes the elementary charge e, the Lorentz force is 
ZeBv. Equating the magnitudes of the two forces, we obtain 


= ZeBv 

and therefore 


Since mv equals the momentum p of the particle, 


(F-1) may 

also be written as 



Cosmic Rays 

This equation, which states that the magnetic rigidity equals the 
ratio of the momentum to the charge of the particle, was derived 
from classical mechanics. (The expression used for the centrifugal 
force is valid only if v «C c.) It turns out (although I shall omit the 
proof) that Eq. (F-2) is relativistically correct; that is, it is correct 
for velocities approaching the velocity of light. [This is not true 

We saw in Appendix D that for relativistic velocities the 
kinetic energy E approaches the product pc. In this case, Eq. 
(F-2) yields 

From the definition of the electron volt (Appendix E) it follows 
that the ratio E/e is the energy of the particle measured in electron 
volts. (A particle of charge e accelerated by a potential difference V 
acquires a kinetic energy E = eV; hence V = E/e.) Thus, writing 
E ev for the energy in electron volts, we have 

cBR = % < F - 3 > 

Equation (F-3) is numerically correct if all quantities are 
measured in the same system of units. We have measured potential 
differences in volts, and the volt belongs to the mks system of 
units. In this system, the unit of length (to be used for R) is the 
meter and the unit of magnetic field strength (to be used for B) is 
the weber per square meter. For consistency, the velocity of light c 
must be measured in meters per second and therefore has the 
approximate numerical value of 3 X 10 s . However, it is customary 
to measure B in gauss rather than in webers per square meter 
(1 gauss = 10~ 4 weber/m 2 ). Also, R is often measured in centi- 
meters rather than meters (1 cm = 10 - ' meter). The propor- 
tionality constant between E cv /Z and BR then becomes 3 X 10 s X 
10-* X 10 -2 = 300, and we arrive at the useful formula cited in 
Chap. 5: 

300 BgauM i? cm = —£■ 


The neutron and 

the structure of 

atomic nuclei 



The neutron is a particle with almost exactly the same mass as the 
proton, but with no electric charge. When it was discovered experi- 
mentally by James Chadwick in 1932, the neutron provided the 
answer to a question about the structure of atomic nuclei that had 
puzzled physicists for many years. In the early part of the century 
it had become clear that the masses of all atomic nuclei were 
approximately whole-number multiples of the proton mass. How- 
ever, the electric charges of nuclei were less than they would have 
been if nuclei consisted entirely of protons. For example, experi- 
ments had shown the helium nucleus to weigh about as much as 
four protons, although its charge was equal to that of only two 

Physicists considered the possibility that atomic nuclei con- 
tained both protons and electrons. If, for example, the helium 
nucleus were made of four protons and two electrons, the negative 
charge of the latter would cancel that of two protons, leaving a net 



Cosmic Rays 

positive charge of two elementary charges. Tliis idea, however, met 
with great difficulties. To mention one, the uncertainly principle, 
a basic law of quantum mechanics, stipulates that a particle obliged 
to remain in a volume with linear dimensions r cannot be at rest, 
but must possess a momentum of the order of h/r (where h is 
Planck's constant, equal to 6.61 X 10 -27 erg sec). 

From measurements of nuclear radii, it was possible to com- 
pute the momentum, and therefore the energy, of the hypothetical 
electrons in nuclei. This energy turned out to be greater than 100 
MeV. It was very hard to accept this conclusion, because electrons 
with such large kinetic energy could not possibly remain confined 
within a nucleus. 1 Moreover, the electrons emitted as rays from 
radioactive nuclei had much lower energies, ranging from a fraction 
of one MeV to a few MeV at most. 

After the discovery of the neutron, it became clear that nuclei 
consisted of protons and neutrons, and not protons and electrons. 
A helium nucleus, then, contains two protons and two neutrons; the 
nucleus of oxygen contains eight protons and eight neutrons. The 
total number of protons and neutrons in a nucleus is the mass 
number A, and the number of protons (equal to the number of 
positive electric charges of the nucleus) is the charge number Z. 
Thus, for a helium nucleus, A = 4, Z = 2. The same information is 
frequently given in a more convenient form: 2 He 4 , where the sub- 
script represents the charge number and the superscript the mass 
number. By this convention an oxygen nucleus is g 16 (that is, 
A = 16,Z= 8). 

Not all nuclei with the same charge number Z have the same 
mass number A. Atoms whose nuclei have different mass numbers 
but the same charge number (and therefore the same number of 
electrons around each nucleus) are called isotopes. Isotopes 
behave in almost identical manners chemically, because chemical 
reactions depend mainly on the configuration of the electron cloud 
and are very little influenced by the nuclear mass. Therefore, 
isotopes occupy the same place in the periodic table of elements 
(hence their name, derived from the Greek words for "same" and 

1 The s:i n ii- difficulty did not arise in the case of protons because, for a 
given momentum, the energy of a proton is much lower than that of an electron. 

The neutron and the structure of atomic nuclei — appendix G 249 

"place"). For example, magnesium, the twelfth element in the 
periodic table (Z = 12), has three isotopes, with mass numbers 21, 
25, and 26, respectively ( 12 Mg", , s Mg 25 , i S Mg 26 ). 

If it is true that nuclei contain only protons and neutrons, 
the electrons emitted in the decay of radioactive nuclei must be 
produced at the moment of the decay. But, because of the law of 
conservation of electric charge, the birth of an electron must be 
accompanied by the creation of a positive charge equal in magni- 
tude to that of the negative electron. According to the accepted 
view, the positive charge appears when one of the neutrons in the 
nucleus changes into a proton. Free neutrons (that is, neutrons that 
are not part of a nucleus) have been observed to decay into protons 
and electrons (see Appendix H). 

It should be noted that all nuclei weigh slightly less than their 
component protons and neutrons would weigh separately. Because 
protons and neutrons in a nucleus attract one another with large 
forces (known as nuclear forces), a certain amount of energy AE 
is required to separate a given nucleus into its component particles. 
Conversely, an equal amount of energy is released when free pro- 
tons and neutrons unite to form a nucleus. This release of energy is 
accompanied by the loss of a certain amount of mass AM given by 
Einstein's equation 

AE = AM c 2 (G -1 ) 

One of the most important experimental data of nuclear 
physics is the mass defect AM of the various nuclei. The first 
accurate determinations of mass defects were made by the English 
physicist F. W. Aston in the first quarter of this century. His 
results are shown, in graphical form, in Fig. G-l. Plotted on the 
horizontal axis are mass numhers A. Plotted on the vertical axis 
are the energies AM c 2 (corresponding to the observed mass 
defects of various nuclei) divided by the corresponding mass 
numbers A and expressed in MeV. The quantity AM c"-/A is the 
average energy per particle that would be needed to split a nucleus 
into the individual protons and neutrons of which it is formed. 
Figure G-l shows that this energy has a maximum value (of about 
8.5 MeV) for nuclei of mass number between 50 and 60. These 
nuclei, therefore, are the most stable nuclei occurring in nature. 


Cosmic Rays 




80 100 120 140 

Atomic mass number, A 

Fig. G-l Energy per particle (AM c*/A) required to split a nucleus into the 
protons and neutrons of which it is formed, plotted against the mass number A 
of the nucleus. The dots are experimental points obtained from the measure- 
ments of Aston. 

The neutrino 



The discovery of the neutrino, unlike that of the neutron, was the 
result of theoretical reasoning rather than experiment. It had been 
known for some time that whereas the a particles emitted by the 
nuclei of a given radioactive clement all have the same energy (or, 
at most, a few sharply defined energies), particles have a variety 
of energies ranging from practically zero to some maximum energy 
characteristic of the radioactive nucleus from which they came. 
Physicists were greatly puzzled by this fact, for it was very hard to 
believe that the decay of a particular nucleus could occur by a 
large number of different decay processes, each releasing a different 
amount of energy. The difficulty was so great that some physicists 
even began to wonder whether decay violated the principle of 
conservation of energy. 

In 1931 the theoretical physicist Wolfgang Pauli found a way 
out of this difficulty by postulating that, in decay, not one par- 
ticle (a negative electron) but two particles were emitted simul- 
taneously. One was the electron, the other was an "invisible" 
particle, or a particle that passed through matter without pro- 
ducing any disturbance by which it could be detected. According 
to this argument, the decay of a particular nucleus would always 
release the same amount of energy. The energy could, however, be 
shared in different proportions between the electron and the 



Cosmic Rays 

hypothetical invisible particle. The particle in question could not 
carry any electric charge; for it would have betrayed its presence 
by a trail of ions along its path. Also, in order to account for the 
known experimental facts about /3 rays, it was necessary to assume 
that the mass of the invisible particle was small compared to that 
of an electron, or perhaps was zero. 

Pauli's idea was taken up and developed by Enrico Fermi, who 
showed in 1934 that the hypothetical particle made it possible to 
explain the properties of /3 decay not only qualitatively but 
quantitatively as well. The particle had first been given the name 
"neutron." After Chadwick's discovery, the name was changed to 
neutrino (Italian for "the little neutral one"). Neutrinos interact 
so weakly with matter that physicists did not succeed in detecting 
them until some 20 years later, after nuclear reactors had made 
available exceedingly strong sources of neutrinos. (The experiment 
was performed by F. Reines, C. L. Cowan, and others in 1955.) 

Free neutrons n decay spontaneously into protons p, electrons 
e~, and neutrinos v. 

>p + e- + v 


with a mean life of about 16.6 minutes. (The decay was first ob- 
served in 1950.) The neutron mass m„ is slightly greater than the 
proton mass m p — the difference between the corresponding rest 
energies (nine* — m p c 2 ) being approximately 1.3 MeV. This energy 
difference accounts for the rest energy and kinetic energy of the 
electron and for the energy of the neutrino emitted in the decay. 
The decay of a neutron bound in a nucleus may be accelerated, 
retarded, or precluded altogether by the close interaction of the 
neutron with other neutrons and protons in the nucleus. As noted 
in Appendix G, the decay of neutrons in radioactive nuclei is the 
fundamental process responsible for the /3-ray emission. In nuclei 
that are not /3 active, on the other hand, neutrons are stable ; that 
is, they do not decay. 

While free protons are stable (because they have a rest energy 
smaller than that of neutrons), protons belonging to nuclei of some 
artificially produced elements decay spontaneously into a neutron, 
a positron, and a neutrino (or, rather, an antineutrino v\ see 
Appendix I): 

The neutrino — appendix II 


p ^ n + e + + - v (H-2) 

This decay process is responsible for the so-called positive 


Neutrinos are produced not only in the (spontaneous or 
induced) decay of neutrons and protons but also in the spontaneous 
decay of other elementary particles such as the y. meson and the 
■k meson (see Chaps. 8 and 9). However, it was shown recently 
(by L. Lederman, M. Schwartz, and R. J. Steinberg in 1962) that 
the neutrinos associated with the decay of t mesons into p. mesons 
(v„) are different from those emitted, together with electrons, when 
a neutron changes into a proton, or vice versa (»,). Thus there exist 
in nature at least two different kinds of neutrinos (and anti- 

Elementary particles 


A list of the known elementary particles and a description of some 
of their properties appear in Table 1-1. In this table the particles 
are subdivided into four families: 

1. The pholon (7) — the quantum of electromagnetic radia- 
tion, that is, the quantum of the electromagnetic field of forces. 

2. The neutrinos [of which, as noted in Appendix H, there 
exist two different kinds (v. and »„)], the electrons [negative (<r) and 
positive (e+)], and the n mesons [negative Or) and positive (m + )1- 
Neutrinos, electrons, and ^ mesons are collectively called leplons. 

3. Pi mesons [positive (*■+), negative (*-), and neutral (x )] 
and K mesons [positive (K + ), negative (K~), and neutral (K ]). Pi 
mesons may be regarded as the quanta of the nuclear field of forces 
— the cohesive forces between protons and neutrons that hold 
nuclei together. K mesons are some sort of "heavy" pi mesons. 

4. The nucleons [that is, protons (p) and neutrons (n)] and 
the various hyperons [lambda particles (A), sigma particles (2), and 
xi particles (H)], some of which are neutral and some charged. 
Collectively these particles are called baryons. 

All members of the third and fourth families interact strongly 
with one another by means of the so-called nuclear forces. No such 
interaction occurs between two leptons, or between a lepton and a 
particle of another family. 



Cosmic Rays 

All electrically charged elementary particles have charges 
equal to the positive or negative elementary charge (1.6 X lO - " 
coulomb). In the table, the symbol + in the column under 
"charge" indicates a positive elementary charge, the symbol - a 
negative elementary charge, and the symbol no charge. The law 
of conservation of charge states that in an isolated system the total 
number of positive elementary charges minus the total number of 
negative elementary charges always remains constant. For exam- 
ple, the creation of a negative electron in the materialization 
process of a photon is accompanied by the creation of a positive 
electron (positron) so that the net charge is zero both before and 
after the event. 

To each particle there corresponds an antiparticle with iden- 
tical mass, so that when a particle and an antiparticle meet, they 
annihilate each other. The energy released (equal to the combined 
mass times the square of the velocity of light) appears in the form 
of photons or of fast-moving particles (for example, the annihilation 
of a negative electron with a positive electron gives rise to two 
photons; the annihilation of a proton with an antiproton results 
in the production of jr mesons). Conversely, all particle-anti- 
particle pairs may be created by materialization of energy. 

When a particle is unstable, its antiparticle is also unstable 
and the two have identical mean lives. Moreover, the decay 
products of one are the antiparticles of the decay products of the 
other. If a particle is positively charged, its antiparticle is nega- 
tively charged, and vice versa. 

In the lepton and baryon families, a neutral particle is physi- 
cally different from its antiparticle. Thus a neutron (n) is different 
from an antineutron (n) because the neutron decay gives rise to a 
(positively charged) proton and a negative electron, while the 
antineutron decay gives rise to a (negatively charged) antiproton 
and a positive electron. In the other two families, however, neutral 
particles are identical to their antiparticles (for example, the *" 
meson is its own antiparticle). 

Leptons and baryons obey two conservation laws that for- 
mally resemble the law of conservation of electric charge: the 
conservation of leptons and the conservation of baryons. 

Elementary particles — appendix I 


The first law states that, in an isolated system, the total 
number of leptons, minus the total number of antileptons, remains 
constant. For example, in the decay process 

n~ — ► e~ + v. + v„ 
there is one lepton (m~) before the event and there are two leptons 
(c - and v„) and one antilepton (v.) after the event. 

Similarly, the second law states that, in an isolated system, 
the total number of baryons minus the total number of anti- 
baryons remains constant. For example, in the simultaneous anni- 
hilation of a proton and an antiproton, there are one baryon and 
one antibaryon before the event and there are no baryons or anti- 
baryons after the event. 

In principle, the choice as to which particle of a pair is called 
a particle and which an antiparticle is an arbitrary one. However, 
it is customary to regard the negative electron and the proton 
(which are among the building blocks of ordinary matter) as 
particles. Consequently the character of all other baryons and 
leptons is uniquely determined. For example, the ordinary neutron 
is a baryon, because it decays into a proton and two particles that 
do not belong to the baryon family. 

Conservation of baryons is responsible for the stability of 
matter; it prevents a proton from changing into, say, five positrons 
and four electrons, a process not forbidden by the conservation of 
energy and electric charge. 

No conservation law similar to the conservation of leptons and 
baryons exists for the other two families of particles. 












_u _o _o _« 

2 -a j 3 

CO t/J C/5 Cfl 

72 yi 

I + 

+ + 

+ + 



I + 

NO \© 

O — 
CI (N 










o+ I 

+ >1 % 
++++ + + 


w i*fc t-i. i* 

++ ++ ++ ++ 


+ * % + a 'a + „ V, V % Vat + f + * + k 'a 'a UVl 

X X 



VO t- t- 

s E 
< 3 


V v 


b< * 










a — m 



= ■ 






oo+ I 



■fi «. 

+ + 

+ + 

V % V> \ % t, >■ >-T. \ \ V V l» \ X, \ T. V \ 
+++++ +++++++++++++++ 
a. b o.<5.« e.«SKi a, 6 «.« s ec«4M<M 





















ll X 






: X 






V V 



s s 


C f P3M 

o©+ I I + ©o I + 

04 C-l CO 

co co c-i 






vo *> CO CO 
tnmifl in 




c * 



B A. 



/I (mass number), 248 
Absorbable particles, 109 
Absorption, 23 
Absorption curve, 24, 25, 78 

exponential, 24, 25 
Air showers, 180, 221 

arrival directions of, 190 

core of, 182 

determination of propagation direc- 
tion of, 187 

discovery of. 180, 181 

largest observed, 192 

lateral spread of (density of parti- 
cles as function of distance from 
center), 188-190 

longitudinal spread of, 187 

production mechanism of, 182, 183 

size spectrum of (frequency as func- 
tion of size), 190, 191 
a particles, 14 

in primary cosmic rays, 167 
a rays (see a particles) 
Alfven, Hannes, 196 
Alvarez, Luis W., 69 
Anderson, Carl D., 79, 103, 141, 158 
Anderson, Kinsey, 208 
Annihilation, of particles with anti- 
particles, 256 

of positrons with electrons, 85 
Antineutrinos, 140, 252 
Antineutrons, 160, 256 


Antineutrons, discovery of, 160 
Antiparticles, 256 

Dirac's theory of, 84, 85 

(See also Antineutrinos; Antineu- 
trons; Antiprotons; Positrons) 
Antiprotons, 160 

discovery of, 160 
Aruki, Gentaro, 126 
Argus, Project, 199 
Aston, F. W., 249 
Atmospheric depth, 11, 233 

as function of altitude, 177 
Atoms, 3, 14 
Auger, Pierre, 181 

Aurorae (northern lights), 59, 212, 

Bachelet, F., 211 

Balata, P., 211 

Baryons, 255 

Bassi, Pietro, 187 

Bccquerel, Henri, 5 

Belcnky, S. Z., 95 

Bernardini, Gilberto, 43, 119, 168 

decay, 113, 249. 251 

Fermi's theory of, 252 

of neutron, 249, 252 
particles (fi rays), 15 
Betatron, 220, 226 
Bethe, Hans A., 90, 92, 143n. 




BeV, 17, 243 

Bhabha, Homi J., 95 

Blackett, P. M. S., 50, 83, 85, 114, 116 

Blau, Marietta, 146 

Bohr, Niels, 116 

Bothe, Walther, 30, 59 

BR (see Magnetic rigidity) 

Bradt, Hans L., 165 

Bremsstrahlung (see Radiation loss) 

Bridge, Herbert, 212 

Brode, Robert B., 107 

Brown, R. H.. 135, 152. 158 

Bush, Vannevar, 66 

Butler, C. C, 156, 157 

Camerini, U., 135, 152, 158 

Carbon 14(,C"), 161 

Carbon dating, 161 

Carlson, J. F., 95 

Cascade effect in electric discharge, 

Cascade process in production of 

showers, 95 
Cathode rays, 15 
Chadwick, James, 15, 85 
Chakrabarty, S. K... 95 
Chamberlain, Owen, 160 
Chaminade, R., 124 
Charge number (2), 248 
Christofilos, N. C, 199 
Chudakov, A. E., 200 
Clark, George W., 166, 185 
Clay, J., 59 
Cloud chamber, 18 

counter-controlled, 49 

diffusion, 49 

expansion, 18 

multiplate, 130, 149 
Cocconi, Giuseppe, 119 
Coincidence, 34 

chance, 46 
Coincidence circuit, 44 
Compton, A. H., 22. 68, 69, 71, 73, 

Compton effect, 22 

Compton effect, probability of, as 
function of energy, 93 
theory of, 28n. 
Conforto, A. M., 211 
Conservation law, of baryons, 256 
of charge, 256 
of energy, 138 
of leptons, 256 
of momentum, 138 
of parity, 160 
Conversi, Marcello, 124, 127 
Corlin, Axel, 59 
Corpuscular radiation, 15 
Cosmic radiation (see Cosmic rays) 
Cosmic-ray intensity, as function of 
atmospheric depth, 11 
under water, 12 

in vertical direction as function of 
atmospheric depth, 167-169 
Cosmic-ray particles, penetrating, 52, 
soft, 52, 78 
Cosmic-ray stars, 146, 151 
Cosmic-ray telescope, 68 

for measurement of directional in- 
tensity, 173 
Cosmic rays, 8 
local, 52 

origin of name, 8 
penetration in lead, 45 
primary, 34, 52 

composition of, 167 
energy spectrum of, 173, 174 
flux of, 174 
nature of, 163-166 
origin of, 219, 226, 227 
secondary effects of, 47 
Counter, Geiger-Miiller, 33 
point, 31 
scintillation, 32 
Cowan, C. L., 252 
Crab nebula, 224 
Cross section, nuclear, 153 
Curie, Marie, 5 
Curie, Pierre, 5 



Dc Benedetti, Sergio, 70 
Decay products, of K mesons, 159 
of A particles, 159 
of m mesons, 141 
of *• mesons, 138-140 
of t° mesons, 142 
of Z particles, 159 
Dirac, P. A. M., 28, 84 
Direction, allowed, 62 

forbidden, 62 
Directional intensity, definition of, 
of electrons as function of atmos- 
pheric depth, 176 
of m mesons as function of atmos- 
pheric depth, 176 
of nuclear-active particles as func- 
tion of atmospheric depth, 176 
of primary cosmic rays as function 
of energy, 174 
Discharge, electric, 32 
Disintegration products (see Decay 

Earl, James, 166 

Earth, distance from sun (astronom- 
ical unit), 221 

radius of, 58, 221 
East-west effect, 65, 173 

discovery of, 69, 70 
Einstein, Albert, 239 
Electric discharge, 32 
Electromagnetic radiations, 15 
Electron volt (eV), 17, 243 
Electrons, 4, 14, 16 

in local radiation, 171, 172, 175, 
176, 178 

mass of, 16 

positive (see Positrons) 

in primary radiation, 166 

rest energy of, 244 
Electroscope, 2 

gold-leaf, 3 

self-recording, 10 

Wulfs, 6 

Elementary charge, 256 
Elementary particles, 156 

families of, 255 

properties of, 258, 259 
Eleven-year cycle, of cosmic-ray in- 
tensity, 216, 217 

of solar activity, 215 
Emulsion (see Nuclear emulsion) 
Emulsion chamber, 31 
Energy, kinetic (see Kinetic energy) 

rest (see Rest energy) 
Energy loss, 91 

by ionization, 92 

by radiation, 91, 92 
Energy spectrum of primary cosmic 

rays, 173, 174 
Equivalence principle, 29, 240 
Euler, H.. 114 
eV (electron volt), 17, 243 
Expansion phase, 19 
Exponential absorption, 24, 25 

Fermi, Enrico, 65, 113, 226, 252 
Forbush, Scott E., 210 
Forbush decreases, 210 

in interplanetary space, 214 

origin of, 214 
Fowler, P. H., 135, 152, 158 
Freon, A., 124 
Furry, Wendell H., 95 

Galaxy, 228 
y rays, 15 

Garbasso, Antonio, 43 
Gauss (unit), 55, 56, 246 
Geiger, Hans, 31, 32 
Geiger-Miiller counter, 33 
Geomagnetic axis, 54 
Geomagnetic effects, 59-69 
Geomagnetic equator, 54 
Geomagnetic field, 53, 54 
Ginzburg, V. L., 167, 223 
G-M counter, 33 


Gockel, A.. 5 
Grain density, 131 

as function of range, 132 
Grams per square centimeter, 233 
Greisen, Kenneth I., 122 
Guiding center, 196 
Guiding line, 196 

Half-life, 113n. 

Hall, David B., 119 

Heavy mesons (K mesons), 158, 159 

in local radiation, 172, 175 
Heavy nuclei in primary cosmic rays, 

Heisenberg, Werner, 114, 146 
Heitler, W., 90, 92, 95, 152 
Helium nucleus, 16 
Hess, Victor F., 1 
Hilberry, Norman, 116 
Hoag, J. Barton, 116 
Hyperons, 159 

in local radiation, 172, 175 

Infrared rays, 15 
Interaction, nuclear, 145, 255 

weak, 129 
Ion, 4 

Ion density, 19 
dependence, on charge, 19 

on kinetic energy, 21 

on magnetic rigidity (BR), 80, 81 

on velocity, 19 
Ion drift, 32 
Ion pair, 7 
Ionization, 4 

by corpuscular rays, 17-20 
by photons, 20-23 
Ionization density (see Ion density) 
Ionization loss, 92 
Isocosms, 71 
Isotopes, 248 

Janossy, L., 147 
Johnson, Thomas H., 69 
Joliot-Curie, Frederic, 85 


Joliot-Curie, Irene, 85 
Joule expressed in eV, 244 

K mesons (see Heavy mesons) 

KeV, 243 

Kinetic energy, 16, 239 
classical expression of, 239 
as function of velocity, 241 
relativistic expression of, 240 

King, D. T., 152 

Klein, Oscar, 28 

Kohlhbrster, W., 7, 30, 59 

KorfT, Serge, 161 

Krassowsky, V. I., 200 

Kraushaar, William, 166 

Kuhlenkampff, C., 114 

A particles, 159 
Landau, Lev D., 95 
Latitude effect, 58, 59, 67, 71, 

variation with altitude, 73-75 
Lattes, C. M. G., 133, 134 
Lederman, L., 253 
Lee, T. D., 160 
Leighton, Robert B., 141 
Lcmaitre, Georges E., 66 
Lcprincc-Ringuet, Louis, 156 
Lcptons, 255 
Libby, Willard F., 161 
Light, velocity of, 16 
Light nuclei in primary cosmic rays, 

Ijght year, 225 
Linsley, John, 191 
Liouville theorem, 65 
Local radiation, 52 

nature of, 109 
Ixigarithmic scale, 237 
Longitude effect, 71 
Longitudinal drift, 196 
Lorcntz, Hcndrik, 55 
Ix)rentz force, 55 


Magnetic dipolc, 53 

of earth, 53, 54 
Magnetic field, at earth's equator, 58 
in galaxy, 230, 231 
in interplanetary space, 214, 215, 

217, 221 
of sun, 214 
Magnetic (lux, 196 
Magnetic lens, 127, 138 
Magnetic lines of force, 54, 145 
Magnetic rigidity (B/?), 55 
computation of, 245 
as function of kinetic energy, 56 
Magnetic storms, 196, 210 
main phase of, 210, 213 
origin of, 213 

sudden commencement of, 210, 213 
Magnetic trapping, 194, 199 
Magnetic tubes of force, 195 
Magnetosphere, 213 
Marini, G., 211 
Marshack, Robert E., 143n. 
Mass, 16 

of electron, 16 

of elementary particles, 258, 259 
of helium nucleus, 16 
Mass defect, 29, 240, 249 
Mass determination, from ionization 
density and magnetic rigidity, 
from magnetic rigidity and range, 

from range and grain density, 133 
Mass number (A), 248 
Mass per unit area, 233 
Materialization (see Pair production) 
Maze, R., 124 
Mean free path, 24 
of a particles in atmosphere, 174, 

of primary protons in atmosphere. 
Mean life, 113 

dependence on speed, 117 

of elementary particles, 258, 259 


Mean life, of m mesons, 118, 122, 124 
in motion, 119 
of t mesons, 136, 137 
of x° mesons, 142, 143 
Medium nuclei in primary cosmic 

rays, 167 
Meridian plane, 62 
Meson, 107 

theoretical prediction of, 110 
(See also K mesons; n mesons; *• 
MeV, 17, 243 
Meyer, Peter, 166 
Micron, 143 

Millikan, Robert A... 8, 28, 59, 73 
Minimum-ionizing particles, 20 
Mirror points, 196 
Molecules, 3 
Momentum, 241 

classical expression of, 242 
conservation of, 138 
as function of velocity, 241, 242 
relativistic expression of, 242 
n mesons, 107 

absorption of, in air, 116-118 
decay of, 113 
detected electronically, 121-124 
observed, in cloud chambers, 119, 

observed in nuclear emulsions, 
decay curve of, 123 
from decay of t mesons, 135 
discovery of, 107 
in local radiation, 171, 172, 178 
mass of, 108 

mean life of, 118, 122, 124 
nuclear capture of, 126-129 
Muirhead, H., 134, 135, 158 
Miiller, W., 32 
Multiple production, 146, 147 

Neddermeyer, Seth H., 103 
Neher, Victor H., 119, 217 



Nereson, Norris G., 122 
Neutrinos, 113, 140, 251 

associated with electrons, 253 

associated with n mesons, 253 

direct detection of, 252 

origin of name, 252 

from supernovae, 222n. 
Neutrons, 15, 160, 247 

decay of, 252 

in local radiation, 172, 175 

mean life of, 252 
Ney, Edward P., 165 
Nielsen, W. M., 119 
Nishina, Yoshio, 28 
Nonionizing link, 84 
Northern lights (aurorae). 59, 212, 

Nuclear-active particles, 153 

nature of, 155, 175 

variation with atmospheric depth, 
Nuclear capture, 126, 136, 137 
Nuclear emulsion, 130-133, 149 

stripped, 131 
Nuclear forces, 112, 249, 255 
Nuclear interactions, 145, 255 
Nuclei, 14 

in primary cosmic rays, 165, 
Nucleons, 255 

Occhialini, Giuseppe, 43, 50, 83, 85, 

131, 133, 134, 136 
Oppenheimer, J. R., 91, 95, 141 

1'nir production, 85, 91 

dependence on Z, 92 

probability as function of energy, 
Pancini, Ettore, 127 
Parity, conservation of, 160 
Pauli, Wolfgang. 113, 251 
Penetrating particles, 52 

nature of, 109 

Penetrating showers, 148, 150 
detector of, 148 

variation with altitude, 153, 154 
Perkins, D. H., 136, 151 
Peters, Bernard, 165 
Pfotzer, Georg, 167 
Photons, 16 
in local radiation, 171, 172, 175, 

176, 178 
in primary cosmic rays, 166 
Photosphere, 212 
x mesons, 135 
decay of, 133-135 
discovery of, 133-135 
in local radiation, 171, 172, 175 
mass of, 137 
mean life of, 136, 137 
neutral (see «■" mesons) 
nuclear capture of, 136 
as quanta of nuclear force, 135, 136 
x° mesons, 141-143 
decay of, 142 
discovery of, 142 
mass of, 142 
mean life of, 142, 143 
Piccioni, Oreste, 119, 124, 127 
Planck's constant, 248 
Plasma, 212 
Point counter, 31 
Polar cap absorption, 208-210 
Pomerantz, Martin A., 119 
Positive electrons (see Positrons) 
Positrons, 71 
discovery of, 79-83 
production by 7 rays, 86 
theoretical prediction of, 84, 85 
Potential difference, electric (volt- 
age). 243 
Powell, C. F., 131, 133-137, 152. 156, 

Primary cosmic radiation (see Cosmic 

rays, primary) 
Protons, 14 
in local radiation, 172, 175 
in primary cosmic rays, 165-167 



Protons, rest energy of, 244 
from solar flares, 206 

Radiation, 15 

corpuscular, 15 

electromagnetic, 15 
Radiation belt (Van Allen belt), 193 

composition of, 200 

discovery of, 197-199 

inner, 201 

intensity distribution of, 201, 202 

origin of, 202, 203 

outer, 201 

variations of, accompanying mag- 
netic storms, 212, 213 
Radiation loss, 91 

as function of kinetic energy, 92 
Radio bursts, 224 
Radio noises, 224 
Radioactivity, 5 
Radium, 5 
Range, 23, 133 

as function of magnetic rigidity, 
Rasetti, Franco, 119, 126 
Recoil electron, 23 
Regener, Erich, 10 
Rcines, F., 252 
Relativistic particle, 16, 241 
Relativistic velocity, 16, 241 
Resolving time, 180 
Rest energy, 240 

of electron, 244 

of proton, 244 
Right-hand rule, 56 
Ring current, 196 
Riometcrs. 208 
Ritson, D. If., 135, 158 
Roberts. G. E., 119 
Rochester, George D-, 156, 157 
Roentgen, Wilhelm, 5 
Rutherford, Ernest, 31 

Scarsi, Livio, 191 

Scattering, in nuclear emulsion, 131 

Scattering, of shower particles, 183 

Schcin, Marcel, 156, 164 

Schwartz, M., 253 

Scintillation, 31 

Scintillation counter, 32 

Segre, Emilio, 160 

Serber, Robert, 95 

Seriff, A. J., 141 

Shklovsky, I. S., 225 

Shower, 50 

air (see Air showers) 
penetrating (see Penetrating show- 
Shower curve, 88, 89 

theoretical, 97 
Shower-producting radiation, 90 

nature of, 101 
Shower theory, 94 
2 particles, 159 

Simpson, John A., 161, 200, 207 
Singer, Fred, 196 
Skobeltzyn, D., 47 
Snyder, Hartland S., 95 
Soft component (soft particles), 52, 78 

nature of, 99 
Solar activity, effect on cosmic-ray 
intensity, 210, 216, 217 
effect on radiation belt, 201 
eleven-year cycle of, 215 
Solar corona, 212 
Solar flare, 212, 224 
Solar-flare particles, 206 

trapping of, in solar system, 209 
Solar wind, 212 
Steinberg, R. J., 253 
Stevenson, E. C, 105 
Stonner, Carl, 59, 194 
Stormer cone, 62 

aperture of, as function of mag- 
netic rigidity, 64 
Street, J. C, 105, 149 
Supernovae, 222 

remnants of, 224 
Swann, W. F. G., 226 
Synchrotron, 223 



Synchrotron radiation, 223 
from Crab nebula, 225 
as origin of radio noises, 224 

Syrovatsky, S. I., 167 

Tamaki, H., 143n. 

Tamm, Igor E., 95 

Thompson, Robert W., 119 

Thomson, J. J., 4 

Tinlot, John, 153 

Tomonaga, Sin-itiro, 126, 143n. 

Trajectory, allowed, 61 

bounded, 62 

equatorial, 65 

forbidden, 61 
Trapped particles, 194 

spectrum of, 200 

Ultragammaslrahlung, 27 
Ultraviolet rays, 15 
Uncertainty principle, 248 

Vallarta, Manuel S., 66 
Van Allen, James A., 197 

Van Allen belt (see Radiation belt) 
Velocity of light, 16 
Vcrnov, S. N., 200 
Vogt, Rochus, 166 
von Weiszacker, C, 102 

Wambacher, H., 146 
Wataghin, Glob, 146, 147 
Weak interaction, 129 
Weber per square meter, 246 
Williams, E. J., 102, 119 
Wilson, C. T. R., 7, 18 
Winkler, J. R., 200 
Wu, C. S., 160 
Wulf, Thomas, 5 

X-rays, 5, 15 

Yang, C. M., 160 

Yukawa, Hideki, 112 

Yukawa's theory, 112, 113, 135, 136 

Z (charge number), 248 

539.? r.nsmic Ravs 
RO by Bruno Rossi 

Date Due 

Divine Word Seminary 

Stutlent Library 

Perrysburg, Ohio 

(Wf r.. 

I 1 

(Continued from flap) 

ics, and then went to England with a Fellowship 
of the Society for the Protection of Science and 
Learning, working at Professor Blackett's lab- 
oratory in Manchester. In the summer of 1939 
he came to this country at the invitation of 
Professor Arthur Compton, as a Research As- 
sociate at The University of Chicago. The fol- 
lowing year he was appointed Associate Profes- 
sor of Physics at Cornell University, where he 
stayed until 1943 when he joined the staff of 
the Los Alamos Laboratory. Since 1946 he has 
been Professor of Physics at the Massachu- 
setts Institute of Technology. Professor Rossi 
is a member of the National Academy of Sci- 
ences, the American Academy of Arts and 
Sciences, the American Philosophical Society, 
the Accademia dei Lincei, and the Society of 
Sigma Xi. He is also a member of the Space 
Science Board of the National Academy of 

Professor Rossi is the author of a book on 
"High-Energy Particles," of a book on "Optics" 
and the coauthor of a book on "Ionization 
Chambers and Counters." He is the author or 
coauthor of many papers on the subject of 
cosmic rays. His many achievements and dis- 
coveries in this area are recounted in the pages 
of this book. 

Other McGRAW-HILL Books 

NUCLEAR PHYSICS. An Introduction 

By W. E. BURCHAM, University of Birmingham, England. 739 pages, $12.00 

The author describes the facts and theories of nuclear physics from an experimental 
standpoint. The first part of the book deals with natural radioactivity and with those 
results of the wave mechanical theory of atoms which seem important in describing 
their nuclei. The' second part gives a fairly detailed description of the techniques of 
nuclear physics. The third and fourth parts are devoted to discussions of nuclear prop- 
erties and reactions, and the final section outlines our present knowledge of the 
neutron-proton interaction important to the understanding of the properties of com- 
plex nuclei. 


By ARTHUR BEISER, formerly of New York University. 384 pages, $7.95 

This rigorous, clear, and logically developed treatment of modern physics is dis- 
tinguished by its well-integrated progression from relatively and quantum theory 
through the atom, molecule, and nucleus. It delves into quantum theory early with 
thorough discussions of quantum-mechanical ideas and formulas. The viewpoint is 
contemporary, and the subject matter is introduced with a minimum of engineering 
applications and historical detail. 


By DAVID PARK, Williams College. 575 pages, $10.75 

This book, written with the undergraduate specifically in mind, is uniquely organized 
so that the course could be taught in either of two ways: the first half develops the 
general formalism of the theory, making use of as few and as definite assumptions as 
possible, while the second part gives extensive applications to atomic and nuclear and 
solid state theory and to statistical mechanics. All the theory could be given first, and 
then the applications, starting with Chapter I and proceeding to the end. Or the theory 
could be taught along with the applications, since the chapters in each part are set 
up to parallel each other. 


By P. T. MATTHEWS, Imperial College, London. 170 pages, $5.00 

Offers a sound introduction, providing a step-by-step guide to the basic concepts of 
quantum mechanics. Throughout, mathematics is kept to a minimum, but the funda- 
mental ideas are introduced in such a way that the student should be able to cope 
with more sophisticated discussions of the foundations of the general theory, or to 
more detailed treatments of its special branches. 

McGraw-Hill Book Company 

330 West 42nd Street/ New York, N. Y. 10036